talloc_string_sub2: Don't return NULL if src is empty.
[Samba.git] / source / lib / util_str.c
blob5f26cc80f8c99e6ebc503d2e72576ab94ea076af
1 /*
2 Unix SMB/CIFS implementation.
3 Samba utility functions
5 Copyright (C) Andrew Tridgell 1992-2001
6 Copyright (C) Simo Sorce 2001-2002
7 Copyright (C) Martin Pool 2003
8 Copyright (C) James Peach 2006
9 Copyright (C) Jeremy Allison 1992-2007
11 This program is free software; you can redistribute it and/or modify
12 it under the terms of the GNU General Public License as published by
13 the Free Software Foundation; either version 3 of the License, or
14 (at your option) any later version.
16 This program is distributed in the hope that it will be useful,
17 but WITHOUT ANY WARRANTY; without even the implied warranty of
18 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19 GNU General Public License for more details.
21 You should have received a copy of the GNU General Public License
22 along with this program. If not, see <http://www.gnu.org/licenses/>.
25 #include "includes.h"
27 char toupper_ascii_fast_table[128] = {
28 0x0, 0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7, 0x8, 0x9, 0xa, 0xb, 0xc, 0xd, 0xe, 0xf,
29 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f,
30 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f,
31 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f,
32 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f,
33 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x5a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f,
34 0x60, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f,
35 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x5a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f
38 /**
39 * @file
40 * @brief String utilities.
41 **/
43 static bool next_token_internal_talloc(TALLOC_CTX *ctx,
44 const char **ptr,
45 char **pp_buff,
46 const char *sep,
47 bool ltrim)
49 char *s;
50 char *saved_s;
51 char *pbuf;
52 bool quoted;
53 size_t len=1;
55 *pp_buff = NULL;
56 if (!ptr) {
57 return(false);
60 s = (char *)*ptr;
62 /* default to simple separators */
63 if (!sep) {
64 sep = " \t\n\r";
67 /* find the first non sep char, if left-trimming is requested */
68 if (ltrim) {
69 while (*s && strchr_m(sep,*s)) {
70 s++;
74 /* nothing left? */
75 if (!*s) {
76 return false;
79 /* When restarting we need to go from here. */
80 saved_s = s;
82 /* Work out the length needed. */
83 for (quoted = false; *s &&
84 (quoted || !strchr_m(sep,*s)); s++) {
85 if (*s == '\"') {
86 quoted = !quoted;
87 } else {
88 len++;
92 /* We started with len = 1 so we have space for the nul. */
93 *pp_buff = TALLOC_ARRAY(ctx, char, len);
94 if (!*pp_buff) {
95 return false;
98 /* copy over the token */
99 pbuf = *pp_buff;
100 s = saved_s;
101 for (quoted = false; *s &&
102 (quoted || !strchr_m(sep,*s)); s++) {
103 if ( *s == '\"' ) {
104 quoted = !quoted;
105 } else {
106 *pbuf++ = *s;
110 *ptr = (*s) ? s+1 : s;
111 *pbuf = 0;
113 return true;
116 #if 0
118 * Get the next token from a string, return false if none found. Handles
119 * double-quotes. This version trims leading separator characters before
120 * looking for a token.
122 bool next_token(const char **ptr, char *buff, const char *sep, size_t bufsize)
124 return next_token_internal(ptr, buff, sep, bufsize, true);
126 #endif
128 bool next_token_talloc(TALLOC_CTX *ctx,
129 const char **ptr,
130 char **pp_buff,
131 const char *sep)
133 return next_token_internal_talloc(ctx, ptr, pp_buff, sep, true);
137 * Get the next token from a string, return false if none found. Handles
138 * double-quotes. This version does not trim leading separator characters
139 * before looking for a token.
142 bool next_token_no_ltrim_talloc(TALLOC_CTX *ctx,
143 const char **ptr,
144 char **pp_buff,
145 const char *sep)
147 return next_token_internal_talloc(ctx, ptr, pp_buff, sep, false);
151 * Case insensitive string compararison.
153 * iconv does not directly give us a way to compare strings in
154 * arbitrary unix character sets -- all we can is convert and then
155 * compare. This is expensive.
157 * As an optimization, we do a first pass that considers only the
158 * prefix of the strings that is entirely 7-bit. Within this, we
159 * check whether they have the same value.
161 * Hopefully this will often give the answer without needing to copy.
162 * In particular it should speed comparisons to literal ascii strings
163 * or comparisons of strings that are "obviously" different.
165 * If we find a non-ascii character we fall back to converting via
166 * iconv.
168 * This should never be slower than convering the whole thing, and
169 * often faster.
171 * A different optimization would be to compare for bitwise equality
172 * in the binary encoding. (It would be possible thought hairy to do
173 * both simultaneously.) But in that case if they turn out to be
174 * different, we'd need to restart the whole thing.
176 * Even better is to implement strcasecmp for each encoding and use a
177 * function pointer.
179 int StrCaseCmp(const char *s, const char *t)
182 const char *ps, *pt;
183 size_t size;
184 smb_ucs2_t *buffer_s, *buffer_t;
185 int ret;
187 for (ps = s, pt = t; ; ps++, pt++) {
188 char us, ut;
190 if (!*ps && !*pt)
191 return 0; /* both ended */
192 else if (!*ps)
193 return -1; /* s is a prefix */
194 else if (!*pt)
195 return +1; /* t is a prefix */
196 else if ((*ps & 0x80) || (*pt & 0x80))
197 /* not ascii anymore, do it the hard way
198 * from here on in */
199 break;
201 us = toupper_ascii_fast(*ps);
202 ut = toupper_ascii_fast(*pt);
203 if (us == ut)
204 continue;
205 else if (us < ut)
206 return -1;
207 else if (us > ut)
208 return +1;
211 if (!push_ucs2_allocate(&buffer_s, ps, &size)) {
212 return strcmp(ps, pt);
213 /* Not quite the right answer, but finding the right one
214 under this failure case is expensive, and it's pretty
215 close */
218 if (!push_ucs2_allocate(&buffer_t, pt, &size)) {
219 SAFE_FREE(buffer_s);
220 return strcmp(ps, pt);
221 /* Not quite the right answer, but finding the right one
222 under this failure case is expensive, and it's pretty
223 close */
226 ret = strcasecmp_w(buffer_s, buffer_t);
227 SAFE_FREE(buffer_s);
228 SAFE_FREE(buffer_t);
229 return ret;
234 Case insensitive string compararison, length limited.
236 int StrnCaseCmp(const char *s, const char *t, size_t len)
238 size_t n = 0;
239 const char *ps, *pt;
240 size_t size;
241 smb_ucs2_t *buffer_s, *buffer_t;
242 int ret;
244 for (ps = s, pt = t; n < len ; ps++, pt++, n++) {
245 char us, ut;
247 if (!*ps && !*pt)
248 return 0; /* both ended */
249 else if (!*ps)
250 return -1; /* s is a prefix */
251 else if (!*pt)
252 return +1; /* t is a prefix */
253 else if ((*ps & 0x80) || (*pt & 0x80))
254 /* not ascii anymore, do it the
255 * hard way from here on in */
256 break;
258 us = toupper_ascii_fast(*ps);
259 ut = toupper_ascii_fast(*pt);
260 if (us == ut)
261 continue;
262 else if (us < ut)
263 return -1;
264 else if (us > ut)
265 return +1;
268 if (n == len) {
269 return 0;
272 if (!push_ucs2_allocate(&buffer_s, ps, &size)) {
273 return strncmp(ps, pt, len-n);
274 /* Not quite the right answer, but finding the right one
275 under this failure case is expensive,
276 and it's pretty close */
279 if (!push_ucs2_allocate(&buffer_t, pt, &size)) {
280 SAFE_FREE(buffer_s);
281 return strncmp(ps, pt, len-n);
282 /* Not quite the right answer, but finding the right one
283 under this failure case is expensive,
284 and it's pretty close */
287 ret = strncasecmp_w(buffer_s, buffer_t, len-n);
288 SAFE_FREE(buffer_s);
289 SAFE_FREE(buffer_t);
290 return ret;
294 * Compare 2 strings.
296 * @note The comparison is case-insensitive.
298 bool strequal(const char *s1, const char *s2)
300 if (s1 == s2)
301 return(true);
302 if (!s1 || !s2)
303 return(false);
305 return(StrCaseCmp(s1,s2)==0);
309 * Compare 2 strings up to and including the nth char.
311 * @note The comparison is case-insensitive.
313 bool strnequal(const char *s1,const char *s2,size_t n)
315 if (s1 == s2)
316 return(true);
317 if (!s1 || !s2 || !n)
318 return(false);
320 return(StrnCaseCmp(s1,s2,n)==0);
324 Compare 2 strings (case sensitive).
327 bool strcsequal(const char *s1,const char *s2)
329 if (s1 == s2)
330 return(true);
331 if (!s1 || !s2)
332 return(false);
334 return(strcmp(s1,s2)==0);
338 Do a case-insensitive, whitespace-ignoring string compare.
341 int strwicmp(const char *psz1, const char *psz2)
343 /* if BOTH strings are NULL, return TRUE, if ONE is NULL return */
344 /* appropriate value. */
345 if (psz1 == psz2)
346 return (0);
347 else if (psz1 == NULL)
348 return (-1);
349 else if (psz2 == NULL)
350 return (1);
352 /* sync the strings on first non-whitespace */
353 while (1) {
354 while (isspace((int)*psz1))
355 psz1++;
356 while (isspace((int)*psz2))
357 psz2++;
358 if (toupper_ascii(*psz1) != toupper_ascii(*psz2) ||
359 *psz1 == '\0' || *psz2 == '\0')
360 break;
361 psz1++;
362 psz2++;
364 return (*psz1 - *psz2);
368 Convert a string to "normal" form.
371 void strnorm(char *s, int case_default)
373 if (case_default == CASE_UPPER)
374 strupper_m(s);
375 else
376 strlower_m(s);
380 Check if a string is in "normal" case.
383 bool strisnormal(const char *s, int case_default)
385 if (case_default == CASE_UPPER)
386 return(!strhaslower(s));
388 return(!strhasupper(s));
393 String replace.
394 NOTE: oldc and newc must be 7 bit characters
396 void string_replace( char *s, char oldc, char newc )
398 char *p;
400 /* this is quite a common operation, so we want it to be
401 fast. We optimise for the ascii case, knowing that all our
402 supported multi-byte character sets are ascii-compatible
403 (ie. they match for the first 128 chars) */
405 for (p = s; *p; p++) {
406 if (*p & 0x80) /* mb string - slow path. */
407 break;
408 if (*p == oldc) {
409 *p = newc;
413 if (!*p)
414 return;
416 /* Slow (mb) path. */
417 #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
418 /* With compose characters we must restart from the beginning. JRA. */
419 p = s;
420 #endif
422 while (*p) {
423 size_t c_size;
424 next_codepoint(p, &c_size);
426 if (c_size == 1) {
427 if (*p == oldc) {
428 *p = newc;
431 p += c_size;
436 * Skip past some strings in a buffer - old version - no checks.
437 * **/
439 char *push_skip_string(char *buf)
441 buf += strlen(buf) + 1;
442 return(buf);
446 Skip past a string in a buffer. Buffer may not be
447 null terminated. end_ptr points to the first byte after
448 then end of the buffer.
451 char *skip_string(const char *base, size_t len, char *buf)
453 const char *end_ptr = base + len;
455 if (end_ptr < base || !base || !buf || buf >= end_ptr) {
456 return NULL;
459 /* Skip the string */
460 while (*buf) {
461 buf++;
462 if (buf >= end_ptr) {
463 return NULL;
466 /* Skip the '\0' */
467 buf++;
468 return buf;
472 Count the number of characters in a string. Normally this will
473 be the same as the number of bytes in a string for single byte strings,
474 but will be different for multibyte.
477 size_t str_charnum(const char *s)
479 size_t ret, converted_size;
480 smb_ucs2_t *tmpbuf2 = NULL;
481 if (!push_ucs2_allocate(&tmpbuf2, s, &converted_size)) {
482 return 0;
484 ret = strlen_w(tmpbuf2);
485 SAFE_FREE(tmpbuf2);
486 return ret;
490 Count the number of characters in a string. Normally this will
491 be the same as the number of bytes in a string for single byte strings,
492 but will be different for multibyte.
495 size_t str_ascii_charnum(const char *s)
497 size_t ret, converted_size;
498 char *tmpbuf2 = NULL;
499 if (!push_ascii_allocate(&tmpbuf2, s, &converted_size)) {
500 return 0;
502 ret = strlen(tmpbuf2);
503 SAFE_FREE(tmpbuf2);
504 return ret;
507 bool trim_char(char *s,char cfront,char cback)
509 bool ret = false;
510 char *ep;
511 char *fp = s;
513 /* Ignore null or empty strings. */
514 if (!s || (s[0] == '\0'))
515 return false;
517 if (cfront) {
518 while (*fp && *fp == cfront)
519 fp++;
520 if (!*fp) {
521 /* We ate the string. */
522 s[0] = '\0';
523 return true;
525 if (fp != s)
526 ret = true;
529 ep = fp + strlen(fp) - 1;
530 if (cback) {
531 /* Attempt ascii only. Bail for mb strings. */
532 while ((ep >= fp) && (*ep == cback)) {
533 ret = true;
534 if ((ep > fp) && (((unsigned char)ep[-1]) & 0x80)) {
535 /* Could be mb... bail back to tim_string. */
536 char fs[2], bs[2];
537 if (cfront) {
538 fs[0] = cfront;
539 fs[1] = '\0';
541 bs[0] = cback;
542 bs[1] = '\0';
543 return trim_string(s, cfront ? fs : NULL, bs);
544 } else {
545 ep--;
548 if (ep < fp) {
549 /* We ate the string. */
550 s[0] = '\0';
551 return true;
555 ep[1] = '\0';
556 memmove(s, fp, ep-fp+2);
557 return ret;
561 Trim the specified elements off the front and back of a string.
564 bool trim_string(char *s,const char *front,const char *back)
566 bool ret = false;
567 size_t front_len;
568 size_t back_len;
569 size_t len;
571 /* Ignore null or empty strings. */
572 if (!s || (s[0] == '\0'))
573 return false;
575 front_len = front? strlen(front) : 0;
576 back_len = back? strlen(back) : 0;
578 len = strlen(s);
580 if (front_len) {
581 while (len && strncmp(s, front, front_len)==0) {
582 /* Must use memmove here as src & dest can
583 * easily overlap. Found by valgrind. JRA. */
584 memmove(s, s+front_len, (len-front_len)+1);
585 len -= front_len;
586 ret=true;
590 if (back_len) {
591 while ((len >= back_len) &&
592 strncmp(s+len-back_len,back,back_len)==0) {
593 s[len-back_len]='\0';
594 len -= back_len;
595 ret=true;
598 return ret;
602 Does a string have any uppercase chars in it?
605 bool strhasupper(const char *s)
607 smb_ucs2_t *tmp, *p;
608 bool ret;
609 size_t converted_size;
611 if (!push_ucs2_allocate(&tmp, s, &converted_size)) {
612 return false;
615 for(p = tmp; *p != 0; p++) {
616 if(isupper_w(*p)) {
617 break;
621 ret = (*p != 0);
622 SAFE_FREE(tmp);
623 return ret;
627 Does a string have any lowercase chars in it?
630 bool strhaslower(const char *s)
632 smb_ucs2_t *tmp, *p;
633 bool ret;
634 size_t converted_size;
636 if (!push_ucs2_allocate(&tmp, s, &converted_size)) {
637 return false;
640 for(p = tmp; *p != 0; p++) {
641 if(islower_w(*p)) {
642 break;
646 ret = (*p != 0);
647 SAFE_FREE(tmp);
648 return ret;
652 Find the number of 'c' chars in a string
655 size_t count_chars(const char *s,char c)
657 smb_ucs2_t *ptr;
658 int count;
659 smb_ucs2_t *alloc_tmpbuf = NULL;
660 size_t converted_size;
662 if (!push_ucs2_allocate(&alloc_tmpbuf, s, &converted_size)) {
663 return 0;
666 for(count=0,ptr=alloc_tmpbuf;*ptr;ptr++)
667 if(*ptr==UCS2_CHAR(c))
668 count++;
670 SAFE_FREE(alloc_tmpbuf);
671 return(count);
675 Safe string copy into a known length string. maxlength does not
676 include the terminating zero.
679 char *safe_strcpy_fn(const char *fn,
680 int line,
681 char *dest,
682 const char *src,
683 size_t maxlength)
685 size_t len;
687 if (!dest) {
688 DEBUG(0,("ERROR: NULL dest in safe_strcpy, "
689 "called from [%s][%d]\n", fn, line));
690 return NULL;
693 #ifdef DEVELOPER
694 clobber_region(fn,line,dest, maxlength+1);
695 #endif
697 if (!src) {
698 *dest = 0;
699 return dest;
702 len = strnlen(src, maxlength+1);
704 if (len > maxlength) {
705 DEBUG(0,("ERROR: string overflow by "
706 "%lu (%lu - %lu) in safe_strcpy [%.50s]\n",
707 (unsigned long)(len-maxlength), (unsigned long)len,
708 (unsigned long)maxlength, src));
709 len = maxlength;
712 memmove(dest, src, len);
713 dest[len] = 0;
714 return dest;
718 Safe string cat into a string. maxlength does not
719 include the terminating zero.
721 char *safe_strcat_fn(const char *fn,
722 int line,
723 char *dest,
724 const char *src,
725 size_t maxlength)
727 size_t src_len, dest_len;
729 if (!dest) {
730 DEBUG(0,("ERROR: NULL dest in safe_strcat, "
731 "called from [%s][%d]\n", fn, line));
732 return NULL;
735 if (!src)
736 return dest;
738 src_len = strnlen(src, maxlength + 1);
739 dest_len = strnlen(dest, maxlength + 1);
741 #ifdef DEVELOPER
742 clobber_region(fn, line, dest + dest_len, maxlength + 1 - dest_len);
743 #endif
745 if (src_len + dest_len > maxlength) {
746 DEBUG(0,("ERROR: string overflow by %d "
747 "in safe_strcat [%.50s]\n",
748 (int)(src_len + dest_len - maxlength), src));
749 if (maxlength > dest_len) {
750 memcpy(&dest[dest_len], src, maxlength - dest_len);
752 dest[maxlength] = 0;
753 return NULL;
756 memcpy(&dest[dest_len], src, src_len);
757 dest[dest_len + src_len] = 0;
758 return dest;
762 Paranoid strcpy into a buffer of given length (includes terminating
763 zero. Strips out all but 'a-Z0-9' and the character in other_safe_chars
764 and replaces with '_'. Deliberately does *NOT* check for multibyte
765 characters. Don't change it !
768 char *alpha_strcpy_fn(const char *fn,
769 int line,
770 char *dest,
771 const char *src,
772 const char *other_safe_chars,
773 size_t maxlength)
775 size_t len, i;
777 #ifdef DEVELOPER
778 clobber_region(fn, line, dest, maxlength);
779 #endif
781 if (!dest) {
782 DEBUG(0,("ERROR: NULL dest in alpha_strcpy, "
783 "called from [%s][%d]\n", fn, line));
784 return NULL;
787 if (!src) {
788 *dest = 0;
789 return dest;
792 len = strlen(src);
793 if (len >= maxlength)
794 len = maxlength - 1;
796 if (!other_safe_chars)
797 other_safe_chars = "";
799 for(i = 0; i < len; i++) {
800 int val = (src[i] & 0xff);
801 if (isupper_ascii(val) || islower_ascii(val) ||
802 isdigit(val) || strchr_m(other_safe_chars, val))
803 dest[i] = src[i];
804 else
805 dest[i] = '_';
808 dest[i] = '\0';
810 return dest;
814 Like strncpy but always null terminates. Make sure there is room!
815 The variable n should always be one less than the available size.
817 char *StrnCpy_fn(const char *fn, int line,char *dest,const char *src,size_t n)
819 char *d = dest;
821 #ifdef DEVELOPER
822 clobber_region(fn, line, dest, n+1);
823 #endif
825 if (!dest) {
826 DEBUG(0,("ERROR: NULL dest in StrnCpy, "
827 "called from [%s][%d]\n", fn, line));
828 return(NULL);
831 if (!src) {
832 *dest = 0;
833 return(dest);
836 while (n-- && (*d = *src)) {
837 d++;
838 src++;
841 *d = 0;
842 return(dest);
845 #if 0
847 Like strncpy but copies up to the character marker. always null terminates.
848 returns a pointer to the character marker in the source string (src).
851 static char *strncpyn(char *dest, const char *src, size_t n, char c)
853 char *p;
854 size_t str_len;
856 #ifdef DEVELOPER
857 clobber_region(dest, n+1);
858 #endif
859 p = strchr_m(src, c);
860 if (p == NULL) {
861 DEBUG(5, ("strncpyn: separator character (%c) not found\n", c));
862 return NULL;
865 str_len = PTR_DIFF(p, src);
866 strncpy(dest, src, MIN(n, str_len));
867 dest[str_len] = '\0';
869 return p;
871 #endif
874 Routine to get hex characters and turn them into a 16 byte array.
875 the array can be variable length, and any non-hex-numeric
876 characters are skipped. "0xnn" or "0Xnn" is specially catered
877 for.
879 valid examples: "0A5D15"; "0x15, 0x49, 0xa2"; "59\ta9\te3\n"
883 size_t strhex_to_str(char *buf, size_t buf_len, const char *strhex, size_t strhex_len)
885 size_t i;
886 size_t num_chars = 0;
887 unsigned char lonybble, hinybble;
888 const char *hexchars = "0123456789ABCDEF";
889 char *p1 = NULL, *p2 = NULL;
891 for (i = 0; i < strhex_len && strhex[i] != 0; i++) {
892 if (strnequal(hexchars, "0x", 2)) {
893 i++; /* skip two chars */
894 continue;
897 if (!(p1 = strchr_m(hexchars, toupper_ascii(strhex[i]))))
898 break;
900 i++; /* next hex digit */
902 if (!(p2 = strchr_m(hexchars, toupper_ascii(strhex[i]))))
903 break;
905 /* get the two nybbles */
906 hinybble = PTR_DIFF(p1, hexchars);
907 lonybble = PTR_DIFF(p2, hexchars);
909 if (num_chars >= buf_len) {
910 break;
912 buf[num_chars] = (hinybble << 4) | lonybble;
913 num_chars++;
915 p1 = NULL;
916 p2 = NULL;
918 return num_chars;
921 DATA_BLOB strhex_to_data_blob(TALLOC_CTX *mem_ctx, const char *strhex)
923 DATA_BLOB ret_blob;
925 if (mem_ctx != NULL)
926 ret_blob = data_blob_talloc(mem_ctx, NULL, strlen(strhex)/2+1);
927 else
928 ret_blob = data_blob(NULL, strlen(strhex)/2+1);
930 ret_blob.length = strhex_to_str((char*)ret_blob.data,
931 ret_blob.length,
932 strhex,
933 strlen(strhex));
935 return ret_blob;
939 * Routine to print a buffer as HEX digits, into an allocated string.
942 char *hex_encode(TALLOC_CTX *mem_ctx, const unsigned char *buff_in, size_t len)
944 int i;
945 char *hex_buffer;
947 hex_buffer = TALLOC_ARRAY(mem_ctx, char, (len*2)+1);
949 for (i = 0; i < len; i++)
950 slprintf(&hex_buffer[i*2], 3, "%02X", buff_in[i]);
952 return hex_buffer;
956 Check if a string is part of a list.
959 bool in_list(const char *s, const char *list, bool casesensitive)
961 char *tok = NULL;
962 bool ret = false;
963 TALLOC_CTX *frame;
965 if (!list) {
966 return false;
969 frame = talloc_stackframe();
970 while (next_token_talloc(frame, &list, &tok,LIST_SEP)) {
971 if (casesensitive) {
972 if (strcmp(tok,s) == 0) {
973 ret = true;
974 break;
976 } else {
977 if (StrCaseCmp(tok,s) == 0) {
978 ret = true;
979 break;
983 TALLOC_FREE(frame);
984 return ret;
987 /* this is used to prevent lots of mallocs of size 1 */
988 static const char null_string[] = "";
991 Set a string value, allocing the space for the string
994 static bool string_init(char **dest,const char *src)
996 size_t l;
998 if (!src)
999 src = "";
1001 l = strlen(src);
1003 if (l == 0) {
1004 *dest = CONST_DISCARD(char*, null_string);
1005 } else {
1006 (*dest) = SMB_STRDUP(src);
1007 if ((*dest) == NULL) {
1008 DEBUG(0,("Out of memory in string_init\n"));
1009 return false;
1012 return(true);
1016 Free a string value.
1019 void string_free(char **s)
1021 if (!s || !(*s))
1022 return;
1023 if (*s == null_string)
1024 *s = NULL;
1025 SAFE_FREE(*s);
1029 Set a string value, deallocating any existing space, and allocing the space
1030 for the string
1033 bool string_set(char **dest,const char *src)
1035 string_free(dest);
1036 return(string_init(dest,src));
1040 Substitute a string for a pattern in another string. Make sure there is
1041 enough room!
1043 This routine looks for pattern in s and replaces it with
1044 insert. It may do multiple replacements or just one.
1046 Any of " ; ' $ or ` in the insert string are replaced with _
1047 if len==0 then the string cannot be extended. This is different from the old
1048 use of len==0 which was for no length checks to be done.
1051 void string_sub2(char *s,const char *pattern, const char *insert, size_t len,
1052 bool remove_unsafe_characters, bool replace_once,
1053 bool allow_trailing_dollar)
1055 char *p;
1056 ssize_t ls,lp,li, i;
1058 if (!insert || !pattern || !*pattern || !s)
1059 return;
1061 ls = (ssize_t)strlen(s);
1062 lp = (ssize_t)strlen(pattern);
1063 li = (ssize_t)strlen(insert);
1065 if (len == 0)
1066 len = ls + 1; /* len is number of *bytes* */
1068 while (lp <= ls && (p = strstr_m(s,pattern))) {
1069 if (ls + (li-lp) >= len) {
1070 DEBUG(0,("ERROR: string overflow by "
1071 "%d in string_sub(%.50s, %d)\n",
1072 (int)(ls + (li-lp) - len),
1073 pattern, (int)len));
1074 break;
1076 if (li != lp) {
1077 memmove(p+li,p+lp,strlen(p+lp)+1);
1079 for (i=0;i<li;i++) {
1080 switch (insert[i]) {
1081 case '`':
1082 case '"':
1083 case '\'':
1084 case ';':
1085 case '$':
1086 /* allow a trailing $
1087 * (as in machine accounts) */
1088 if (allow_trailing_dollar && (i == li - 1 )) {
1089 p[i] = insert[i];
1090 break;
1092 case '%':
1093 case '\r':
1094 case '\n':
1095 if ( remove_unsafe_characters ) {
1096 p[i] = '_';
1097 /* yes this break should be here
1098 * since we want to fall throw if
1099 * not replacing unsafe chars */
1100 break;
1102 default:
1103 p[i] = insert[i];
1106 s = p + li;
1107 ls += (li-lp);
1109 if (replace_once)
1110 break;
1114 void string_sub_once(char *s, const char *pattern,
1115 const char *insert, size_t len)
1117 string_sub2( s, pattern, insert, len, true, true, false );
1120 void string_sub(char *s,const char *pattern, const char *insert, size_t len)
1122 string_sub2( s, pattern, insert, len, true, false, false );
1125 void fstring_sub(char *s,const char *pattern,const char *insert)
1127 string_sub(s, pattern, insert, sizeof(fstring));
1131 Similar to string_sub2, but it will accept only allocated strings
1132 and may realloc them so pay attention at what you pass on no
1133 pointers inside strings, no const may be passed
1134 as string.
1137 char *realloc_string_sub2(char *string,
1138 const char *pattern,
1139 const char *insert,
1140 bool remove_unsafe_characters,
1141 bool allow_trailing_dollar)
1143 char *p, *in;
1144 char *s;
1145 ssize_t ls,lp,li,ld, i;
1147 if (!insert || !pattern || !*pattern || !string || !*string)
1148 return NULL;
1150 s = string;
1152 in = SMB_STRDUP(insert);
1153 if (!in) {
1154 DEBUG(0, ("realloc_string_sub: out of memory!\n"));
1155 return NULL;
1157 ls = (ssize_t)strlen(s);
1158 lp = (ssize_t)strlen(pattern);
1159 li = (ssize_t)strlen(insert);
1160 ld = li - lp;
1161 for (i=0;i<li;i++) {
1162 switch (in[i]) {
1163 case '`':
1164 case '"':
1165 case '\'':
1166 case ';':
1167 case '$':
1168 /* allow a trailing $
1169 * (as in machine accounts) */
1170 if (allow_trailing_dollar && (i == li - 1 )) {
1171 break;
1173 case '%':
1174 case '\r':
1175 case '\n':
1176 if ( remove_unsafe_characters ) {
1177 in[i] = '_';
1178 break;
1180 default:
1181 /* ok */
1182 break;
1186 while ((p = strstr_m(s,pattern))) {
1187 if (ld > 0) {
1188 int offset = PTR_DIFF(s,string);
1189 string = (char *)SMB_REALLOC(string, ls + ld + 1);
1190 if (!string) {
1191 DEBUG(0, ("realloc_string_sub: "
1192 "out of memory!\n"));
1193 SAFE_FREE(in);
1194 return NULL;
1196 p = string + offset + (p - s);
1198 if (li != lp) {
1199 memmove(p+li,p+lp,strlen(p+lp)+1);
1201 memcpy(p, in, li);
1202 s = p + li;
1203 ls += ld;
1205 SAFE_FREE(in);
1206 return string;
1209 char *realloc_string_sub(char *string,
1210 const char *pattern,
1211 const char *insert)
1213 return realloc_string_sub2(string, pattern, insert, true, false);
1217 * Internal guts of talloc_string_sub and talloc_all_string_sub.
1218 * talloc version of string_sub2.
1221 char *talloc_string_sub2(TALLOC_CTX *mem_ctx, const char *src,
1222 const char *pattern,
1223 const char *insert,
1224 bool remove_unsafe_characters,
1225 bool replace_once,
1226 bool allow_trailing_dollar)
1228 char *p, *in;
1229 char *s;
1230 char *string;
1231 ssize_t ls,lp,li,ld, i;
1233 if (!insert || !pattern || !*pattern || !src) {
1234 return NULL;
1237 string = talloc_strdup(mem_ctx, src);
1238 if (string == NULL) {
1239 DEBUG(0, ("talloc_string_sub2: "
1240 "talloc_strdup failed\n"));
1241 return NULL;
1244 s = string;
1246 in = SMB_STRDUP(insert);
1247 if (!in) {
1248 DEBUG(0, ("talloc_string_sub2: ENOMEM\n"));
1249 return NULL;
1251 ls = (ssize_t)strlen(s);
1252 lp = (ssize_t)strlen(pattern);
1253 li = (ssize_t)strlen(insert);
1254 ld = li - lp;
1256 for (i=0;i<li;i++) {
1257 switch (in[i]) {
1258 case '`':
1259 case '"':
1260 case '\'':
1261 case ';':
1262 case '$':
1263 /* allow a trailing $
1264 * (as in machine accounts) */
1265 if (allow_trailing_dollar && (i == li - 1 )) {
1266 break;
1268 case '%':
1269 case '\r':
1270 case '\n':
1271 if (remove_unsafe_characters) {
1272 in[i] = '_';
1273 break;
1275 default:
1276 /* ok */
1277 break;
1281 while ((p = strstr_m(s,pattern))) {
1282 if (ld > 0) {
1283 int offset = PTR_DIFF(s,string);
1284 string = (char *)TALLOC_REALLOC(mem_ctx, string,
1285 ls + ld + 1);
1286 if (!string) {
1287 DEBUG(0, ("talloc_string_sub: out of "
1288 "memory!\n"));
1289 SAFE_FREE(in);
1290 return NULL;
1292 p = string + offset + (p - s);
1294 if (li != lp) {
1295 memmove(p+li,p+lp,strlen(p+lp)+1);
1297 memcpy(p, in, li);
1298 s = p + li;
1299 ls += ld;
1301 if (replace_once) {
1302 break;
1305 SAFE_FREE(in);
1306 return string;
1309 /* Same as string_sub, but returns a talloc'ed string */
1311 char *talloc_string_sub(TALLOC_CTX *mem_ctx,
1312 const char *src,
1313 const char *pattern,
1314 const char *insert)
1316 return talloc_string_sub2(mem_ctx, src, pattern, insert,
1317 true, false, false);
1321 Similar to string_sub() but allows for any character to be substituted.
1322 Use with caution!
1323 if len==0 then the string cannot be extended. This is different from the old
1324 use of len==0 which was for no length checks to be done.
1327 void all_string_sub(char *s,const char *pattern,const char *insert, size_t len)
1329 char *p;
1330 ssize_t ls,lp,li;
1332 if (!insert || !pattern || !s)
1333 return;
1335 ls = (ssize_t)strlen(s);
1336 lp = (ssize_t)strlen(pattern);
1337 li = (ssize_t)strlen(insert);
1339 if (!*pattern)
1340 return;
1342 if (len == 0)
1343 len = ls + 1; /* len is number of *bytes* */
1345 while (lp <= ls && (p = strstr_m(s,pattern))) {
1346 if (ls + (li-lp) >= len) {
1347 DEBUG(0,("ERROR: string overflow by "
1348 "%d in all_string_sub(%.50s, %d)\n",
1349 (int)(ls + (li-lp) - len),
1350 pattern, (int)len));
1351 break;
1353 if (li != lp) {
1354 memmove(p+li,p+lp,strlen(p+lp)+1);
1356 memcpy(p, insert, li);
1357 s = p + li;
1358 ls += (li-lp);
1362 char *talloc_all_string_sub(TALLOC_CTX *ctx,
1363 const char *src,
1364 const char *pattern,
1365 const char *insert)
1367 return talloc_string_sub2(ctx, src, pattern, insert,
1368 false, false, false);
1372 Write an octal as a string.
1375 char *octal_string(int i)
1377 char *result;
1378 if (i == -1) {
1379 result = talloc_strdup(talloc_tos(), "-1");
1381 else {
1382 result = talloc_asprintf(talloc_tos(), "0%o", i);
1384 SMB_ASSERT(result != NULL);
1385 return result;
1390 Truncate a string at a specified length.
1393 char *string_truncate(char *s, unsigned int length)
1395 if (s && strlen(s) > length)
1396 s[length] = 0;
1397 return s;
1401 Strchr and strrchr_m are very hard to do on general multi-byte strings.
1402 We convert via ucs2 for now.
1405 char *strchr_m(const char *src, char c)
1407 smb_ucs2_t *ws = NULL;
1408 char *s2 = NULL;
1409 smb_ucs2_t *p;
1410 const char *s;
1411 char *ret;
1412 size_t converted_size;
1414 /* characters below 0x3F are guaranteed to not appear in
1415 non-initial position in multi-byte charsets */
1416 if ((c & 0xC0) == 0) {
1417 return strchr(src, c);
1420 /* this is quite a common operation, so we want it to be
1421 fast. We optimise for the ascii case, knowing that all our
1422 supported multi-byte character sets are ascii-compatible
1423 (ie. they match for the first 128 chars) */
1425 for (s = src; *s && !(((unsigned char)s[0]) & 0x80); s++) {
1426 if (*s == c)
1427 return (char *)s;
1430 if (!*s)
1431 return NULL;
1433 #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
1434 /* With compose characters we must restart from the beginning. JRA. */
1435 s = src;
1436 #endif
1438 if (!push_ucs2_allocate(&ws, s, &converted_size)) {
1439 /* Wrong answer, but what can we do... */
1440 return strchr(src, c);
1442 p = strchr_w(ws, UCS2_CHAR(c));
1443 if (!p) {
1444 SAFE_FREE(ws);
1445 return NULL;
1447 *p = 0;
1448 if (!pull_ucs2_allocate(&s2, ws, &converted_size)) {
1449 SAFE_FREE(ws);
1450 /* Wrong answer, but what can we do... */
1451 return strchr(src, c);
1453 ret = (char *)(s+strlen(s2));
1454 SAFE_FREE(ws);
1455 SAFE_FREE(s2);
1456 return ret;
1459 char *strrchr_m(const char *s, char c)
1461 /* characters below 0x3F are guaranteed to not appear in
1462 non-initial position in multi-byte charsets */
1463 if ((c & 0xC0) == 0) {
1464 return strrchr(s, c);
1467 /* this is quite a common operation, so we want it to be
1468 fast. We optimise for the ascii case, knowing that all our
1469 supported multi-byte character sets are ascii-compatible
1470 (ie. they match for the first 128 chars). Also, in Samba
1471 we only search for ascii characters in 'c' and that
1472 in all mb character sets with a compound character
1473 containing c, if 'c' is not a match at position
1474 p, then p[-1] > 0x7f. JRA. */
1477 size_t len = strlen(s);
1478 const char *cp = s;
1479 bool got_mb = false;
1481 if (len == 0)
1482 return NULL;
1483 cp += (len - 1);
1484 do {
1485 if (c == *cp) {
1486 /* Could be a match. Part of a multibyte ? */
1487 if ((cp > s) &&
1488 (((unsigned char)cp[-1]) & 0x80)) {
1489 /* Yep - go slow :-( */
1490 got_mb = true;
1491 break;
1493 /* No - we have a match ! */
1494 return (char *)cp;
1496 } while (cp-- != s);
1497 if (!got_mb)
1498 return NULL;
1501 /* String contained a non-ascii char. Slow path. */
1503 smb_ucs2_t *ws = NULL;
1504 char *s2 = NULL;
1505 smb_ucs2_t *p;
1506 char *ret;
1507 size_t converted_size;
1509 if (!push_ucs2_allocate(&ws, s, &converted_size)) {
1510 /* Wrong answer, but what can we do. */
1511 return strrchr(s, c);
1513 p = strrchr_w(ws, UCS2_CHAR(c));
1514 if (!p) {
1515 SAFE_FREE(ws);
1516 return NULL;
1518 *p = 0;
1519 if (!pull_ucs2_allocate(&s2, ws, &converted_size)) {
1520 SAFE_FREE(ws);
1521 /* Wrong answer, but what can we do. */
1522 return strrchr(s, c);
1524 ret = (char *)(s+strlen(s2));
1525 SAFE_FREE(ws);
1526 SAFE_FREE(s2);
1527 return ret;
1531 /***********************************************************************
1532 Return the equivalent of doing strrchr 'n' times - always going
1533 backwards.
1534 ***********************************************************************/
1536 char *strnrchr_m(const char *s, char c, unsigned int n)
1538 smb_ucs2_t *ws = NULL;
1539 char *s2 = NULL;
1540 smb_ucs2_t *p;
1541 char *ret;
1542 size_t converted_size;
1544 if (!push_ucs2_allocate(&ws, s, &converted_size)) {
1545 /* Too hard to try and get right. */
1546 return NULL;
1548 p = strnrchr_w(ws, UCS2_CHAR(c), n);
1549 if (!p) {
1550 SAFE_FREE(ws);
1551 return NULL;
1553 *p = 0;
1554 if (!pull_ucs2_allocate(&s2, ws, &converted_size)) {
1555 SAFE_FREE(ws);
1556 /* Too hard to try and get right. */
1557 return NULL;
1559 ret = (char *)(s+strlen(s2));
1560 SAFE_FREE(ws);
1561 SAFE_FREE(s2);
1562 return ret;
1565 /***********************************************************************
1566 strstr_m - We convert via ucs2 for now.
1567 ***********************************************************************/
1569 char *strstr_m(const char *src, const char *findstr)
1571 smb_ucs2_t *p;
1572 smb_ucs2_t *src_w, *find_w;
1573 const char *s;
1574 char *s2;
1575 char *retp;
1577 size_t converted_size, findstr_len = 0;
1579 /* for correctness */
1580 if (!findstr[0]) {
1581 return (char*)src;
1584 /* Samba does single character findstr calls a *lot*. */
1585 if (findstr[1] == '\0')
1586 return strchr_m(src, *findstr);
1588 /* We optimise for the ascii case, knowing that all our
1589 supported multi-byte character sets are ascii-compatible
1590 (ie. they match for the first 128 chars) */
1592 for (s = src; *s && !(((unsigned char)s[0]) & 0x80); s++) {
1593 if (*s == *findstr) {
1594 if (!findstr_len)
1595 findstr_len = strlen(findstr);
1597 if (strncmp(s, findstr, findstr_len) == 0) {
1598 return (char *)s;
1603 if (!*s)
1604 return NULL;
1606 #if 1 /* def BROKEN_UNICODE_COMPOSE_CHARACTERS */
1607 /* 'make check' fails unless we do this */
1609 /* With compose characters we must restart from the beginning. JRA. */
1610 s = src;
1611 #endif
1613 if (!push_ucs2_allocate(&src_w, src, &converted_size)) {
1614 DEBUG(0,("strstr_m: src malloc fail\n"));
1615 return NULL;
1618 if (!push_ucs2_allocate(&find_w, findstr, &converted_size)) {
1619 SAFE_FREE(src_w);
1620 DEBUG(0,("strstr_m: find malloc fail\n"));
1621 return NULL;
1624 p = strstr_w(src_w, find_w);
1626 if (!p) {
1627 SAFE_FREE(src_w);
1628 SAFE_FREE(find_w);
1629 return NULL;
1632 *p = 0;
1633 if (!pull_ucs2_allocate(&s2, src_w, &converted_size)) {
1634 SAFE_FREE(src_w);
1635 SAFE_FREE(find_w);
1636 DEBUG(0,("strstr_m: dest malloc fail\n"));
1637 return NULL;
1639 retp = (char *)(s+strlen(s2));
1640 SAFE_FREE(src_w);
1641 SAFE_FREE(find_w);
1642 SAFE_FREE(s2);
1643 return retp;
1647 Convert a string to lower case.
1650 void strlower_m(char *s)
1652 size_t len;
1653 int errno_save;
1655 /* this is quite a common operation, so we want it to be
1656 fast. We optimise for the ascii case, knowing that all our
1657 supported multi-byte character sets are ascii-compatible
1658 (ie. they match for the first 128 chars) */
1660 while (*s && !(((unsigned char)s[0]) & 0x80)) {
1661 *s = tolower_ascii((unsigned char)*s);
1662 s++;
1665 if (!*s)
1666 return;
1668 /* I assume that lowercased string takes the same number of bytes
1669 * as source string even in UTF-8 encoding. (VIV) */
1670 len = strlen(s) + 1;
1671 errno_save = errno;
1672 errno = 0;
1673 unix_strlower(s,len,s,len);
1674 /* Catch mb conversion errors that may not terminate. */
1675 if (errno)
1676 s[len-1] = '\0';
1677 errno = errno_save;
1681 Convert a string to upper case.
1684 void strupper_m(char *s)
1686 size_t len;
1687 int errno_save;
1689 /* this is quite a common operation, so we want it to be
1690 fast. We optimise for the ascii case, knowing that all our
1691 supported multi-byte character sets are ascii-compatible
1692 (ie. they match for the first 128 chars) */
1694 while (*s && !(((unsigned char)s[0]) & 0x80)) {
1695 *s = toupper_ascii_fast((unsigned char)*s);
1696 s++;
1699 if (!*s)
1700 return;
1702 /* I assume that lowercased string takes the same number of bytes
1703 * as source string even in multibyte encoding. (VIV) */
1704 len = strlen(s) + 1;
1705 errno_save = errno;
1706 errno = 0;
1707 unix_strupper(s,len,s,len);
1708 /* Catch mb conversion errors that may not terminate. */
1709 if (errno)
1710 s[len-1] = '\0';
1711 errno = errno_save;
1715 Count the number of UCS2 characters in a string. Normally this will
1716 be the same as the number of bytes in a string for single byte strings,
1717 but will be different for multibyte.
1720 size_t strlen_m(const char *s)
1722 size_t count = 0;
1724 if (!s) {
1725 return 0;
1728 while (*s && !(((uint8_t)*s) & 0x80)) {
1729 s++;
1730 count++;
1733 if (!*s) {
1734 return count;
1737 while (*s) {
1738 size_t c_size;
1739 codepoint_t c = next_codepoint(s, &c_size);
1740 if (c < 0x10000) {
1741 /* Unicode char fits into 16 bits. */
1742 count += 1;
1743 } else {
1744 /* Double-width unicode char - 32 bits. */
1745 count += 2;
1747 s += c_size;
1750 return count;
1754 Count the number of UCS2 characters in a string including the null
1755 terminator.
1758 size_t strlen_m_term(const char *s)
1760 if (!s) {
1761 return 0;
1763 return strlen_m(s) + 1;
1767 * Weird helper routine for the winreg pipe: If nothing is around, return 0,
1768 * if a string is there, include the terminator.
1771 size_t strlen_m_term_null(const char *s)
1773 size_t len;
1774 if (!s) {
1775 return 0;
1777 len = strlen_m(s);
1778 if (len == 0) {
1779 return 0;
1782 return len+1;
1785 Return a RFC2254 binary string representation of a buffer.
1786 Used in LDAP filters.
1787 Caller must free.
1790 char *binary_string_rfc2254(char *buf, int len)
1792 char *s;
1793 int i, j;
1794 const char *hex = "0123456789ABCDEF";
1795 s = (char *)SMB_MALLOC(len * 3 + 1);
1796 if (!s)
1797 return NULL;
1798 for (j=i=0;i<len;i++) {
1799 s[j] = '\\';
1800 s[j+1] = hex[((unsigned char)buf[i]) >> 4];
1801 s[j+2] = hex[((unsigned char)buf[i]) & 0xF];
1802 j += 3;
1804 s[j] = 0;
1805 return s;
1808 char *binary_string(char *buf, int len)
1810 char *s;
1811 int i, j;
1812 const char *hex = "0123456789ABCDEF";
1813 s = (char *)SMB_MALLOC(len * 2 + 1);
1814 if (!s)
1815 return NULL;
1816 for (j=i=0;i<len;i++) {
1817 s[j] = hex[((unsigned char)buf[i]) >> 4];
1818 s[j+1] = hex[((unsigned char)buf[i]) & 0xF];
1819 j += 2;
1821 s[j] = 0;
1822 return s;
1826 Just a typesafety wrapper for snprintf into a fstring.
1829 int fstr_sprintf(fstring s, const char *fmt, ...)
1831 va_list ap;
1832 int ret;
1834 va_start(ap, fmt);
1835 ret = vsnprintf(s, FSTRING_LEN, fmt, ap);
1836 va_end(ap);
1837 return ret;
1841 List of Strings manipulation functions
1844 #define S_LIST_ABS 16 /* List Allocation Block Size */
1846 char **str_list_make(TALLOC_CTX *mem_ctx, const char *string, const char *sep)
1848 char **list;
1849 const char *str;
1850 char *s;
1851 int num, lsize;
1852 char *tok;
1854 if (!string || !*string)
1855 return NULL;
1857 list = TALLOC_ARRAY(mem_ctx, char *, S_LIST_ABS+1);
1858 if (list == NULL) {
1859 return NULL;
1861 lsize = S_LIST_ABS;
1863 s = talloc_strdup(list, string);
1864 if (s == NULL) {
1865 DEBUG(0,("str_list_make: Unable to allocate memory"));
1866 TALLOC_FREE(list);
1867 return NULL;
1869 if (!sep) sep = LIST_SEP;
1871 num = 0;
1872 str = s;
1874 while (next_token_talloc(list, &str, &tok, sep)) {
1876 if (num == lsize) {
1877 char **tmp;
1879 lsize += S_LIST_ABS;
1881 tmp = TALLOC_REALLOC_ARRAY(mem_ctx, list, char *,
1882 lsize + 1);
1883 if (tmp == NULL) {
1884 DEBUG(0,("str_list_make: "
1885 "Unable to allocate memory"));
1886 TALLOC_FREE(list);
1887 return NULL;
1890 list = tmp;
1892 memset (&list[num], 0,
1893 ((sizeof(char**)) * (S_LIST_ABS +1)));
1896 list[num] = tok;
1897 num += 1;
1900 list[num] = NULL;
1902 TALLOC_FREE(s);
1903 return list;
1906 bool str_list_copy(TALLOC_CTX *mem_ctx, char ***dest, const char **src)
1908 char **list;
1909 int i, num;
1911 *dest = NULL;
1912 if (!src)
1913 return false;
1915 num = 0;
1916 while (src[num] != NULL) {
1917 num += 1;
1920 list = TALLOC_ARRAY(mem_ctx, char *, num+1);
1921 if (list == NULL) {
1922 return false;
1925 for (i=0; i<num; i++) {
1926 list[i] = talloc_strdup(list, src[i]);
1927 if (list[i] == NULL) {
1928 TALLOC_FREE(list);
1929 return false;
1932 list[i] = NULL;
1933 *dest = list;
1934 return true;
1938 * Return true if all the elements of the list match exactly.
1940 bool str_list_compare(char **list1, char **list2)
1942 int num;
1944 if (!list1 || !list2)
1945 return (list1 == list2);
1947 for (num = 0; list1[num]; num++) {
1948 if (!list2[num])
1949 return false;
1950 if (!strcsequal(list1[num], list2[num]))
1951 return false;
1953 if (list2[num])
1954 return false; /* if list2 has more elements than list1 fail */
1956 return true;
1959 /******************************************************************************
1960 *****************************************************************************/
1962 int str_list_count( const char **list )
1964 int i = 0;
1966 if ( ! list )
1967 return 0;
1969 /* count the number of list members */
1971 for ( i=0; *list; i++, list++ );
1973 return i;
1976 /******************************************************************************
1977 version of standard_sub_basic() for string lists; uses talloc_sub_basic()
1978 for the work
1979 *****************************************************************************/
1981 bool str_list_sub_basic( char **list, const char *smb_name,
1982 const char *domain_name )
1984 TALLOC_CTX *ctx = list;
1985 char *s, *tmpstr;
1987 while ( *list ) {
1988 s = *list;
1989 tmpstr = talloc_sub_basic(ctx, smb_name, domain_name, s);
1990 if ( !tmpstr ) {
1991 DEBUG(0,("str_list_sub_basic: "
1992 "alloc_sub_basic() return NULL!\n"));
1993 return false;
1996 TALLOC_FREE(*list);
1997 *list = tmpstr;
1999 list++;
2002 return true;
2005 /******************************************************************************
2006 substritute a specific pattern in a string list
2007 *****************************************************************************/
2009 bool str_list_substitute(char **list, const char *pattern, const char *insert)
2011 char *p, *s, *t;
2012 ssize_t ls, lp, li, ld, i, d;
2014 if (!list)
2015 return false;
2016 if (!pattern)
2017 return false;
2018 if (!insert)
2019 return false;
2021 lp = (ssize_t)strlen(pattern);
2022 li = (ssize_t)strlen(insert);
2023 ld = li -lp;
2025 while (*list) {
2026 s = *list;
2027 ls = (ssize_t)strlen(s);
2029 while ((p = strstr_m(s, pattern))) {
2030 t = *list;
2031 d = p -t;
2032 if (ld) {
2033 t = (char *) SMB_MALLOC(ls +ld +1);
2034 if (!t) {
2035 DEBUG(0,("str_list_substitute: "
2036 "Unable to allocate memory"));
2037 return false;
2039 memcpy(t, *list, d);
2040 memcpy(t +d +li, p +lp, ls -d -lp +1);
2041 SAFE_FREE(*list);
2042 *list = t;
2043 ls += ld;
2044 s = t +d +li;
2047 for (i = 0; i < li; i++) {
2048 switch (insert[i]) {
2049 case '`':
2050 case '"':
2051 case '\'':
2052 case ';':
2053 case '$':
2054 case '%':
2055 case '\r':
2056 case '\n':
2057 t[d +i] = '_';
2058 break;
2059 default:
2060 t[d +i] = insert[i];
2065 list++;
2068 return true;
2072 #define IPSTR_LIST_SEP ","
2073 #define IPSTR_LIST_CHAR ','
2076 * Add ip string representation to ipstr list. Used also
2077 * as part of @function ipstr_list_make
2079 * @param ipstr_list pointer to string containing ip list;
2080 * MUST BE already allocated and IS reallocated if necessary
2081 * @param ipstr_size pointer to current size of ipstr_list (might be changed
2082 * as a result of reallocation)
2083 * @param ip IP address which is to be added to list
2084 * @return pointer to string appended with new ip and possibly
2085 * reallocated to new length
2088 static char *ipstr_list_add(char **ipstr_list, const struct ip_service *service)
2090 char *new_ipstr = NULL;
2091 char addr_buf[INET6_ADDRSTRLEN];
2092 int ret;
2094 /* arguments checking */
2095 if (!ipstr_list || !service) {
2096 return NULL;
2099 print_sockaddr(addr_buf,
2100 sizeof(addr_buf),
2101 &service->ss);
2103 /* attempt to convert ip to a string and append colon separator to it */
2104 if (*ipstr_list) {
2105 if (service->ss.ss_family == AF_INET) {
2106 /* IPv4 */
2107 ret = asprintf(&new_ipstr, "%s%s%s:%d", *ipstr_list,
2108 IPSTR_LIST_SEP, addr_buf,
2109 service->port);
2110 } else {
2111 /* IPv6 */
2112 ret = asprintf(&new_ipstr, "%s%s[%s]:%d", *ipstr_list,
2113 IPSTR_LIST_SEP, addr_buf,
2114 service->port);
2116 SAFE_FREE(*ipstr_list);
2117 } else {
2118 if (service->ss.ss_family == AF_INET) {
2119 /* IPv4 */
2120 ret = asprintf(&new_ipstr, "%s:%d", addr_buf,
2121 service->port);
2122 } else {
2123 /* IPv6 */
2124 ret = asprintf(&new_ipstr, "[%s]:%d", addr_buf,
2125 service->port);
2128 if (ret == -1) {
2129 return NULL;
2131 *ipstr_list = new_ipstr;
2132 return *ipstr_list;
2136 * Allocate and initialise an ipstr list using ip adresses
2137 * passed as arguments.
2139 * @param ipstr_list pointer to string meant to be allocated and set
2140 * @param ip_list array of ip addresses to place in the list
2141 * @param ip_count number of addresses stored in ip_list
2142 * @return pointer to allocated ip string
2145 char *ipstr_list_make(char **ipstr_list,
2146 const struct ip_service *ip_list,
2147 int ip_count)
2149 int i;
2151 /* arguments checking */
2152 if (!ip_list || !ipstr_list) {
2153 return 0;
2156 *ipstr_list = NULL;
2158 /* process ip addresses given as arguments */
2159 for (i = 0; i < ip_count; i++) {
2160 *ipstr_list = ipstr_list_add(ipstr_list, &ip_list[i]);
2163 return (*ipstr_list);
2168 * Parse given ip string list into array of ip addresses
2169 * (as ip_service structures)
2170 * e.g. [IPv6]:port,192.168.1.100:389,192.168.1.78, ...
2172 * @param ipstr ip string list to be parsed
2173 * @param ip_list pointer to array of ip addresses which is
2174 * allocated by this function and must be freed by caller
2175 * @return number of successfully parsed addresses
2178 int ipstr_list_parse(const char *ipstr_list, struct ip_service **ip_list)
2180 TALLOC_CTX *frame;
2181 char *token_str = NULL;
2182 size_t count;
2183 int i;
2185 if (!ipstr_list || !ip_list)
2186 return 0;
2188 count = count_chars(ipstr_list, IPSTR_LIST_CHAR) + 1;
2189 if ( (*ip_list = SMB_MALLOC_ARRAY(struct ip_service, count)) == NULL ) {
2190 DEBUG(0,("ipstr_list_parse: malloc failed for %lu entries\n",
2191 (unsigned long)count));
2192 return 0;
2195 frame = talloc_stackframe();
2196 for ( i=0; next_token_talloc(frame, &ipstr_list, &token_str,
2197 IPSTR_LIST_SEP) && i<count; i++ ) {
2198 char *s = token_str;
2199 char *p = strrchr(token_str, ':');
2201 if (p) {
2202 *p = 0;
2203 (*ip_list)[i].port = atoi(p+1);
2206 /* convert single token to ip address */
2207 if (token_str[0] == '[') {
2208 /* IPv6 address. */
2209 s++;
2210 p = strchr(token_str, ']');
2211 if (!p) {
2212 continue;
2214 *p = '\0';
2216 if (!interpret_string_addr(&(*ip_list)[i].ss,
2218 AI_NUMERICHOST)) {
2219 continue;
2222 TALLOC_FREE(frame);
2223 return count;
2227 * Safely free ip string list
2229 * @param ipstr_list ip string list to be freed
2232 void ipstr_list_free(char* ipstr_list)
2234 SAFE_FREE(ipstr_list);
2238 Unescape a URL encoded string, in place.
2241 void rfc1738_unescape(char *buf)
2243 char *p=buf;
2245 while (p && *p && (p=strchr_m(p,'%'))) {
2246 int c1 = p[1];
2247 int c2 = p[2];
2249 if (c1 >= '0' && c1 <= '9')
2250 c1 = c1 - '0';
2251 else if (c1 >= 'A' && c1 <= 'F')
2252 c1 = 10 + c1 - 'A';
2253 else if (c1 >= 'a' && c1 <= 'f')
2254 c1 = 10 + c1 - 'a';
2255 else {p++; continue;}
2257 if (c2 >= '0' && c2 <= '9')
2258 c2 = c2 - '0';
2259 else if (c2 >= 'A' && c2 <= 'F')
2260 c2 = 10 + c2 - 'A';
2261 else if (c2 >= 'a' && c2 <= 'f')
2262 c2 = 10 + c2 - 'a';
2263 else {p++; continue;}
2265 *p = (c1<<4) | c2;
2267 memmove(p+1, p+3, strlen(p+3)+1);
2268 p++;
2272 static const char b64[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
2275 * Decode a base64 string into a DATA_BLOB - simple and slow algorithm
2277 DATA_BLOB base64_decode_data_blob(const char *s)
2279 int bit_offset, byte_offset, idx, i, n;
2280 DATA_BLOB decoded = data_blob(s, strlen(s)+1);
2281 unsigned char *d = decoded.data;
2282 char *p;
2284 n=i=0;
2286 while (*s && (p=strchr_m(b64,*s))) {
2287 idx = (int)(p - b64);
2288 byte_offset = (i*6)/8;
2289 bit_offset = (i*6)%8;
2290 d[byte_offset] &= ~((1<<(8-bit_offset))-1);
2291 if (bit_offset < 3) {
2292 d[byte_offset] |= (idx << (2-bit_offset));
2293 n = byte_offset+1;
2294 } else {
2295 d[byte_offset] |= (idx >> (bit_offset-2));
2296 d[byte_offset+1] = 0;
2297 d[byte_offset+1] |= (idx << (8-(bit_offset-2))) & 0xFF;
2298 n = byte_offset+2;
2300 s++; i++;
2303 if ((n > 0) && (*s == '=')) {
2304 n -= 1;
2307 /* fix up length */
2308 decoded.length = n;
2309 return decoded;
2313 * Decode a base64 string in-place - wrapper for the above
2315 void base64_decode_inplace(char *s)
2317 DATA_BLOB decoded = base64_decode_data_blob(s);
2319 if ( decoded.length != 0 ) {
2320 memcpy(s, decoded.data, decoded.length);
2322 /* null terminate */
2323 s[decoded.length] = '\0';
2324 } else {
2325 *s = '\0';
2328 data_blob_free(&decoded);
2332 * Encode a base64 string into a talloc()ed string caller to free.
2334 * From SQUID: adopted from http://ftp.sunet.se/pub2/gnu/vm/base64-encode.c
2335 * with adjustments
2338 char *base64_encode_data_blob(TALLOC_CTX *mem_ctx, DATA_BLOB data)
2340 int bits = 0;
2341 int char_count = 0;
2342 size_t out_cnt, len, output_len;
2343 char *result;
2345 if (!data.length || !data.data)
2346 return NULL;
2348 out_cnt = 0;
2349 len = data.length;
2350 output_len = data.length * 2 + 4; /* Account for closing bytes. 4 is
2351 * random but should be enough for
2352 * the = and \0 */
2353 result = TALLOC_ARRAY(mem_ctx, char, output_len); /* get us plenty of space */
2354 SMB_ASSERT(result != NULL);
2356 while (len-- && out_cnt < (data.length * 2) - 5) {
2357 int c = (unsigned char) *(data.data++);
2358 bits += c;
2359 char_count++;
2360 if (char_count == 3) {
2361 result[out_cnt++] = b64[bits >> 18];
2362 result[out_cnt++] = b64[(bits >> 12) & 0x3f];
2363 result[out_cnt++] = b64[(bits >> 6) & 0x3f];
2364 result[out_cnt++] = b64[bits & 0x3f];
2365 bits = 0;
2366 char_count = 0;
2367 } else {
2368 bits <<= 8;
2371 if (char_count != 0) {
2372 bits <<= 16 - (8 * char_count);
2373 result[out_cnt++] = b64[bits >> 18];
2374 result[out_cnt++] = b64[(bits >> 12) & 0x3f];
2375 if (char_count == 1) {
2376 result[out_cnt++] = '=';
2377 result[out_cnt++] = '=';
2378 } else {
2379 result[out_cnt++] = b64[(bits >> 6) & 0x3f];
2380 result[out_cnt++] = '=';
2383 result[out_cnt] = '\0'; /* terminate */
2384 return result;
2387 /* read a SMB_BIG_UINT from a string */
2388 SMB_BIG_UINT STR_TO_SMB_BIG_UINT(const char *nptr, const char **entptr)
2391 SMB_BIG_UINT val = -1;
2392 const char *p = nptr;
2394 if (!p) {
2395 if (entptr) {
2396 *entptr = p;
2398 return val;
2401 while (*p && isspace(*p))
2402 p++;
2404 #ifdef LARGE_SMB_OFF_T
2405 sscanf(p,"%llu",&val);
2406 #else /* LARGE_SMB_OFF_T */
2407 sscanf(p,"%lu",&val);
2408 #endif /* LARGE_SMB_OFF_T */
2409 if (entptr) {
2410 while (*p && isdigit(*p))
2411 p++;
2412 *entptr = p;
2415 return val;
2418 /* Convert a size specification to a count of bytes. We accept the following
2419 * suffixes:
2420 * bytes if there is no suffix
2421 * kK kibibytes
2422 * mM mebibytes
2423 * gG gibibytes
2424 * tT tibibytes
2425 * pP whatever the ISO name for petabytes is
2427 * Returns 0 if the string can't be converted.
2429 SMB_OFF_T conv_str_size(const char * str)
2431 SMB_OFF_T lval;
2432 char * end;
2434 if (str == NULL || *str == '\0') {
2435 return 0;
2438 #ifdef HAVE_STRTOULL
2439 if (sizeof(SMB_OFF_T) == 8) {
2440 lval = strtoull(str, &end, 10 /* base */);
2441 } else {
2442 lval = strtoul(str, &end, 10 /* base */);
2444 #else
2445 lval = strtoul(str, &end, 10 /* base */);
2446 #endif
2448 if (end == NULL || end == str) {
2449 return 0;
2452 if (*end) {
2453 SMB_OFF_T lval_orig = lval;
2455 if (strwicmp(end, "K") == 0) {
2456 lval *= (SMB_OFF_T)1024;
2457 } else if (strwicmp(end, "M") == 0) {
2458 lval *= ((SMB_OFF_T)1024 * (SMB_OFF_T)1024);
2459 } else if (strwicmp(end, "G") == 0) {
2460 lval *= ((SMB_OFF_T)1024 * (SMB_OFF_T)1024 *
2461 (SMB_OFF_T)1024);
2462 } else if (strwicmp(end, "T") == 0) {
2463 lval *= ((SMB_OFF_T)1024 * (SMB_OFF_T)1024 *
2464 (SMB_OFF_T)1024 * (SMB_OFF_T)1024);
2465 } else if (strwicmp(end, "P") == 0) {
2466 lval *= ((SMB_OFF_T)1024 * (SMB_OFF_T)1024 *
2467 (SMB_OFF_T)1024 * (SMB_OFF_T)1024 *
2468 (SMB_OFF_T)1024);
2469 } else {
2470 return 0;
2473 /* Primitive attempt to detect wrapping on platforms with
2474 * 4-byte SMB_OFF_T. It's better to let the caller handle
2475 * a failure than some random number.
2477 if (lval_orig <= lval) {
2478 return 0;
2482 return lval;
2485 void string_append(char **left, const char *right)
2487 int new_len = strlen(right) + 1;
2489 if (*left == NULL) {
2490 *left = (char *)SMB_MALLOC(new_len);
2491 *left[0] = '\0';
2492 } else {
2493 new_len += strlen(*left);
2494 *left = (char *)SMB_REALLOC(*left, new_len);
2497 if (*left == NULL) {
2498 return;
2501 safe_strcat(*left, right, new_len-1);
2504 bool add_string_to_array(TALLOC_CTX *mem_ctx,
2505 const char *str, const char ***strings,
2506 int *num)
2508 char *dup_str = talloc_strdup(mem_ctx, str);
2510 *strings = TALLOC_REALLOC_ARRAY(mem_ctx, *strings,
2511 const char *, (*num)+1);
2513 if ((*strings == NULL) || (dup_str == NULL)) {
2514 *num = 0;
2515 return false;
2518 (*strings)[*num] = dup_str;
2519 *num += 1;
2520 return true;
2523 /* Append an sprintf'ed string. Double buffer size on demand. Usable without
2524 * error checking in between. The indiation that something weird happened is
2525 * string==NULL */
2527 void sprintf_append(TALLOC_CTX *mem_ctx, char **string, ssize_t *len,
2528 size_t *bufsize, const char *fmt, ...)
2530 va_list ap;
2531 char *newstr;
2532 int ret;
2533 bool increased;
2535 /* len<0 is an internal marker that something failed */
2536 if (*len < 0)
2537 goto error;
2539 if (*string == NULL) {
2540 if (*bufsize == 0)
2541 *bufsize = 128;
2543 *string = TALLOC_ARRAY(mem_ctx, char, *bufsize);
2544 if (*string == NULL)
2545 goto error;
2548 va_start(ap, fmt);
2549 ret = vasprintf(&newstr, fmt, ap);
2550 va_end(ap);
2552 if (ret < 0)
2553 goto error;
2555 increased = false;
2557 while ((*len)+ret >= *bufsize) {
2558 increased = true;
2559 *bufsize *= 2;
2560 if (*bufsize >= (1024*1024*256))
2561 goto error;
2564 if (increased) {
2565 *string = TALLOC_REALLOC_ARRAY(mem_ctx, *string, char,
2566 *bufsize);
2567 if (*string == NULL) {
2568 goto error;
2572 StrnCpy((*string)+(*len), newstr, ret);
2573 (*len) += ret;
2574 free(newstr);
2575 return;
2577 error:
2578 *len = -1;
2579 *string = NULL;
2583 * asprintf into a string and strupper_m it after that.
2586 int asprintf_strupper_m(char **strp, const char *fmt, ...)
2588 va_list ap;
2589 char *result;
2590 int ret;
2592 va_start(ap, fmt);
2593 ret = vasprintf(&result, fmt, ap);
2594 va_end(ap);
2596 if (ret == -1)
2597 return -1;
2599 strupper_m(result);
2600 *strp = result;
2601 return ret;
2604 char *talloc_asprintf_strupper_m(TALLOC_CTX *t, const char *fmt, ...)
2606 va_list ap;
2607 char *ret;
2609 va_start(ap, fmt);
2610 ret = talloc_vasprintf(t, fmt, ap);
2611 va_end(ap);
2613 if (ret == NULL) {
2614 return NULL;
2616 strupper_m(ret);
2617 return ret;
2620 char *talloc_asprintf_strlower_m(TALLOC_CTX *t, const char *fmt, ...)
2622 va_list ap;
2623 char *ret;
2625 va_start(ap, fmt);
2626 ret = talloc_vasprintf(t, fmt, ap);
2627 va_end(ap);
2629 if (ret == NULL) {
2630 return NULL;
2632 strlower_m(ret);
2633 return ret;
2638 Returns the substring from src between the first occurrence of
2639 the char "front" and the first occurence of the char "back".
2640 Mallocs the return string which must be freed. Not for use
2641 with wide character strings.
2643 char *sstring_sub(const char *src, char front, char back)
2645 char *temp1, *temp2, *temp3;
2646 ptrdiff_t len;
2648 temp1 = strchr(src, front);
2649 if (temp1 == NULL) return NULL;
2650 temp2 = strchr(src, back);
2651 if (temp2 == NULL) return NULL;
2652 len = temp2 - temp1;
2653 if (len <= 0) return NULL;
2654 temp3 = (char*)SMB_MALLOC(len);
2655 if (temp3 == NULL) {
2656 DEBUG(1,("Malloc failure in sstring_sub\n"));
2657 return NULL;
2659 memcpy(temp3, temp1+1, len-1);
2660 temp3[len-1] = '\0';
2661 return temp3;
2664 /********************************************************************
2665 Check a string for any occurrences of a specified list of invalid
2666 characters.
2667 ********************************************************************/
2669 bool validate_net_name( const char *name,
2670 const char *invalid_chars,
2671 int max_len)
2673 int i;
2675 for ( i=0; i<max_len && name[i]; i++ ) {
2676 /* fail if strchr_m() finds one of the invalid characters */
2677 if ( name[i] && strchr_m( invalid_chars, name[i] ) ) {
2678 return false;
2682 return true;
2687 return the number of bytes occupied by a buffer in ASCII format
2688 the result includes the null termination
2689 limited by 'n' bytes
2691 size_t ascii_len_n(const char *src, size_t n)
2693 size_t len;
2695 len = strnlen(src, n);
2696 if (len+1 <= n) {
2697 len += 1;
2700 return len;
2704 return the number of bytes occupied by a buffer in CH_UTF16 format
2705 the result includes the null termination
2707 size_t utf16_len(const void *buf)
2709 size_t len;
2711 for (len = 0; SVAL(buf,len); len += 2) ;
2713 return len + 2;
2717 return the number of bytes occupied by a buffer in CH_UTF16 format
2718 the result includes the null termination
2719 limited by 'n' bytes
2721 size_t utf16_len_n(const void *src, size_t n)
2723 size_t len;
2725 for (len = 0; (len+2 < n) && SVAL(src, len); len += 2) ;
2727 if (len+2 <= n) {
2728 len += 2;
2731 return len;
2734 /*******************************************************************
2735 Add a shell escape character '\' to any character not in a known list
2736 of characters. UNIX charset format.
2737 *******************************************************************/
2739 #define INCLUDE_LIST "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz_/ \t.,"
2740 #define INSIDE_DQUOTE_LIST "$`\n\"\\"
2742 char *escape_shell_string(const char *src)
2744 size_t srclen = strlen(src);
2745 char *ret = SMB_MALLOC_ARRAY(char, (srclen * 2) + 1);
2746 char *dest = ret;
2747 bool in_s_quote = false;
2748 bool in_d_quote = false;
2749 bool next_escaped = false;
2751 if (!ret) {
2752 return NULL;
2755 while (*src) {
2756 size_t c_size;
2757 codepoint_t c = next_codepoint(src, &c_size);
2759 if (c == INVALID_CODEPOINT) {
2760 SAFE_FREE(ret);
2761 return NULL;
2764 if (c_size > 1) {
2765 memcpy(dest, src, c_size);
2766 src += c_size;
2767 dest += c_size;
2768 next_escaped = false;
2769 continue;
2773 * Deal with backslash escaped state.
2774 * This only lasts for one character.
2777 if (next_escaped) {
2778 *dest++ = *src++;
2779 next_escaped = false;
2780 continue;
2784 * Deal with single quote state. The
2785 * only thing we care about is exiting
2786 * this state.
2789 if (in_s_quote) {
2790 if (*src == '\'') {
2791 in_s_quote = false;
2793 *dest++ = *src++;
2794 continue;
2798 * Deal with double quote state. The most
2799 * complex state. We must cope with \, meaning
2800 * possibly escape next char (depending what it
2801 * is), ", meaning exit this state, and possibly
2802 * add an \ escape to any unprotected character
2803 * (listed in INSIDE_DQUOTE_LIST).
2806 if (in_d_quote) {
2807 if (*src == '\\') {
2809 * Next character might be escaped.
2810 * We have to peek. Inside double
2811 * quotes only INSIDE_DQUOTE_LIST
2812 * characters are escaped by a \.
2815 char nextchar;
2817 c = next_codepoint(&src[1], &c_size);
2818 if (c == INVALID_CODEPOINT) {
2819 SAFE_FREE(ret);
2820 return NULL;
2822 if (c_size > 1) {
2824 * Don't escape the next char.
2825 * Just copy the \.
2827 *dest++ = *src++;
2828 continue;
2831 nextchar = src[1];
2833 if (nextchar && strchr(INSIDE_DQUOTE_LIST,
2834 (int)nextchar)) {
2835 next_escaped = true;
2837 *dest++ = *src++;
2838 continue;
2841 if (*src == '\"') {
2842 /* Exit double quote state. */
2843 in_d_quote = false;
2844 *dest++ = *src++;
2845 continue;
2849 * We know the character isn't \ or ",
2850 * so escape it if it's any of the other
2851 * possible unprotected characters.
2854 if (strchr(INSIDE_DQUOTE_LIST, (int)*src)) {
2855 *dest++ = '\\';
2857 *dest++ = *src++;
2858 continue;
2862 * From here to the end of the loop we're
2863 * not in the single or double quote state.
2866 if (*src == '\\') {
2867 /* Next character must be escaped. */
2868 next_escaped = true;
2869 *dest++ = *src++;
2870 continue;
2873 if (*src == '\'') {
2874 /* Go into single quote state. */
2875 in_s_quote = true;
2876 *dest++ = *src++;
2877 continue;
2880 if (*src == '\"') {
2881 /* Go into double quote state. */
2882 in_d_quote = true;
2883 *dest++ = *src++;
2884 continue;
2887 /* Check if we need to escape the character. */
2889 if (!strchr(INCLUDE_LIST, (int)*src)) {
2890 *dest++ = '\\';
2892 *dest++ = *src++;
2894 *dest++ = '\0';
2895 return ret;