4 Copyright (C) Andrew Tridgell 2004
6 ** NOTE! The following LGPL license applies to the ldb
7 ** library. This does NOT imply that all of Samba is released
10 This library is free software; you can redistribute it and/or
11 modify it under the terms of the GNU Lesser General Public
12 License as published by the Free Software Foundation; either
13 version 3 of the License, or (at your option) any later version.
15 This library is distributed in the hope that it will be useful,
16 but WITHOUT ANY WARRANTY; without even the implied warranty of
17 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
18 Lesser General Public License for more details.
20 You should have received a copy of the GNU Lesser General Public
21 License along with this library; if not, see <http://www.gnu.org/licenses/>.
27 * Component: ldb expression parsing
29 * Description: parse LDAP-like search expressions
31 * Author: Andrew Tridgell
36 - add RFC2254 binary string handling
37 - possibly add ~=, <= and >= handling
38 - expand the test suite
39 - add better parse error handling
44 #include "ldb/include/includes.h"
45 #include "system/locale.h"
47 struct ldb_val
ldb_binary_decode(void *mem_ctx
, const char *str
);
50 a filter is defined by:
51 <filter> ::= '(' <filtercomp> ')'
52 <filtercomp> ::= <and> | <or> | <not> | <simple>
53 <and> ::= '&' <filterlist>
54 <or> ::= '|' <filterlist>
55 <not> ::= '!' <filter>
56 <filterlist> ::= <filter> | <filter> <filterlist>
57 <simple> ::= <attributetype> <filtertype> <attributevalue>
58 <filtertype> ::= '=' | '~=' | '<=' | '>='
62 decode a RFC2254 binary string representation of a buffer.
65 struct ldb_val
ldb_binary_decode(void *mem_ctx
, const char *str
)
69 int slen
= str
?strlen(str
):0;
71 ret
.data
= (uint8_t *)talloc_size(mem_ctx
, slen
+1);
73 if (ret
.data
== NULL
) return ret
;
75 for (i
=j
=0;i
<slen
;i
++) {
78 if (sscanf(&str
[i
+1], "%02X", &c
) != 1) {
79 talloc_free(ret
.data
);
80 memset(&ret
, 0, sizeof(ret
));
83 ((uint8_t *)ret
.data
)[j
++] = c
;
86 ((uint8_t *)ret
.data
)[j
++] = str
[i
];
90 ((uint8_t *)ret
.data
)[j
] = 0;
95 static bool need_encode(unsigned char cval
)
97 if (cval
< 0x20 || cval
> 0x7E || strchr(" *()\\&|!\"", cval
)) {
104 encode a blob as a RFC2254 binary string, escaping any
105 non-printable or '\' characters
107 char *ldb_binary_encode(void *mem_ctx
, struct ldb_val val
)
111 int len
= val
.length
;
112 unsigned char *buf
= val
.data
;
114 for (i
=0;i
<val
.length
;i
++) {
115 if (need_encode(buf
[i
])) {
119 ret
= talloc_array(mem_ctx
, char, len
+1);
120 if (ret
== NULL
) return NULL
;
123 for (i
=0;i
<val
.length
;i
++) {
124 if (need_encode(buf
[i
])) {
125 snprintf(ret
+len
, 4, "\\%02X", buf
[i
]);
138 encode a string as a RFC2254 binary string, escaping any
139 non-printable or '\' characters. This routine is suitable for use
140 in escaping user data in ldap filters.
142 char *ldb_binary_encode_string(void *mem_ctx
, const char *string
)
145 val
.data
= discard_const_p(uint8_t, string
);
146 val
.length
= strlen(string
);
147 return ldb_binary_encode(mem_ctx
, val
);
150 /* find the first matching wildcard */
151 static char *ldb_parse_find_wildcard(char *value
)
154 value
= strpbrk(value
, "\\*");
155 if (value
== NULL
) return NULL
;
157 if (value
[0] == '\\') {
158 if (value
[1] == '\0') return NULL
;
163 if (value
[0] == '*') return value
;
169 /* return a NULL terminated list of binary strings representing the value
170 chunks separated by wildcards that makes the value portion of the filter
172 static struct ldb_val
**ldb_wildcard_decode(void *mem_ctx
, const char *string
)
174 struct ldb_val
**ret
= NULL
;
178 wc
= talloc_strdup(mem_ctx
, string
);
179 if (wc
== NULL
) return NULL
;
183 wc
= ldb_parse_find_wildcard(str
);
193 ret
= talloc_realloc(mem_ctx
, ret
, struct ldb_val
*, val
+ 2);
194 if (ret
== NULL
) return NULL
;
196 ret
[val
] = talloc(mem_ctx
, struct ldb_val
);
197 if (ret
[val
] == NULL
) return NULL
;
199 *(ret
[val
]) = ldb_binary_decode(mem_ctx
, str
);
200 if ((ret
[val
])->data
== NULL
) return NULL
;
212 static struct ldb_parse_tree
*ldb_parse_filter(void *mem_ctx
, const char **s
);
216 parse an extended match
224 the ':dn' part sets the dnAttributes boolean if present
225 the oid sets the rule_id string
228 static struct ldb_parse_tree
*ldb_parse_extended(struct ldb_parse_tree
*ret
,
229 char *attr
, char *value
)
233 ret
->operation
= LDB_OP_EXTENDED
;
234 ret
->u
.extended
.value
= ldb_binary_decode(ret
, value
);
235 if (ret
->u
.extended
.value
.data
== NULL
) goto failed
;
237 p1
= strchr(attr
, ':');
238 if (p1
== NULL
) goto failed
;
239 p2
= strchr(p1
+1, ':');
244 ret
->u
.extended
.attr
= attr
;
245 if (strcmp(p1
+1, "dn") == 0) {
246 ret
->u
.extended
.dnAttributes
= 1;
248 ret
->u
.extended
.rule_id
= talloc_strdup(ret
, p2
+1);
249 if (ret
->u
.extended
.rule_id
== NULL
) goto failed
;
251 ret
->u
.extended
.rule_id
= NULL
;
254 ret
->u
.extended
.dnAttributes
= 0;
255 ret
->u
.extended
.rule_id
= talloc_strdup(ret
, p1
+1);
256 if (ret
->u
.extended
.rule_id
== NULL
) goto failed
;
266 static enum ldb_parse_op
ldb_parse_filtertype(void *mem_ctx
, char **type
, char **value
, const char **s
)
268 enum ldb_parse_op filter
= 0;
269 char *name
, *val
, *k
;
273 /* retrieve attributetype name */
276 while ((isascii(*p
) && isalnum((unsigned char)*p
)) || (*p
== '-')) { /* attribute names can only be alphanums */
280 if (*p
== ':') { /* but extended searches have : and . chars too */
282 if (p
== NULL
) { /* malformed attribute name */
289 while (isspace((unsigned char)*p
)) p
++;
291 if (!strchr("=<>~:", *p
)) {
296 name
= (char *)talloc_memdup(mem_ctx
, t
, t1
- t
+ 1);
297 if (name
== NULL
) return 0;
300 /* retrieve filtertype */
303 filter
= LDB_OP_EQUALITY
;
304 } else if (*(p
+ 1) == '=') {
307 filter
= LDB_OP_LESS
;
311 filter
= LDB_OP_GREATER
;
315 filter
= LDB_OP_APPROX
;
319 filter
= LDB_OP_EXTENDED
;
330 while (isspace((unsigned char)*p
)) p
++;
335 while (*p
&& ((*p
!= ')') || ((*p
== ')') && (*(p
- 1) == '\\')))) p
++;
337 val
= (char *)talloc_memdup(mem_ctx
, t
, p
- t
+ 1);
346 /* remove trailing spaces from value */
347 while ((k
> val
) && (isspace((unsigned char)*(k
- 1)))) k
--;
357 <simple> ::= <attributetype> <filtertype> <attributevalue>
359 static struct ldb_parse_tree
*ldb_parse_simple(void *mem_ctx
, const char **s
)
362 struct ldb_parse_tree
*ret
;
363 enum ldb_parse_op filtertype
;
365 ret
= talloc(mem_ctx
, struct ldb_parse_tree
);
371 filtertype
= ldb_parse_filtertype(ret
, &attr
, &value
, s
);
377 switch (filtertype
) {
380 ret
->operation
= LDB_OP_PRESENT
;
381 ret
->u
.present
.attr
= attr
;
384 case LDB_OP_EQUALITY
:
386 if (strcmp(value
, "*") == 0) {
387 ret
->operation
= LDB_OP_PRESENT
;
388 ret
->u
.present
.attr
= attr
;
392 if (ldb_parse_find_wildcard(value
) != NULL
) {
393 ret
->operation
= LDB_OP_SUBSTRING
;
394 ret
->u
.substring
.attr
= attr
;
395 ret
->u
.substring
.start_with_wildcard
= 0;
396 ret
->u
.substring
.end_with_wildcard
= 0;
397 ret
->u
.substring
.chunks
= ldb_wildcard_decode(ret
, value
);
398 if (ret
->u
.substring
.chunks
== NULL
){
403 ret
->u
.substring
.start_with_wildcard
= 1;
404 if (value
[strlen(value
) - 1] == '*')
405 ret
->u
.substring
.end_with_wildcard
= 1;
411 ret
->operation
= LDB_OP_EQUALITY
;
412 ret
->u
.equality
.attr
= attr
;
413 ret
->u
.equality
.value
= ldb_binary_decode(ret
, value
);
414 if (ret
->u
.equality
.value
.data
== NULL
) {
422 ret
->operation
= LDB_OP_GREATER
;
423 ret
->u
.comparison
.attr
= attr
;
424 ret
->u
.comparison
.value
= ldb_binary_decode(ret
, value
);
425 if (ret
->u
.comparison
.value
.data
== NULL
) {
433 ret
->operation
= LDB_OP_LESS
;
434 ret
->u
.comparison
.attr
= attr
;
435 ret
->u
.comparison
.value
= ldb_binary_decode(ret
, value
);
436 if (ret
->u
.comparison
.value
.data
== NULL
) {
444 ret
->operation
= LDB_OP_APPROX
;
445 ret
->u
.comparison
.attr
= attr
;
446 ret
->u
.comparison
.value
= ldb_binary_decode(ret
, value
);
447 if (ret
->u
.comparison
.value
.data
== NULL
) {
454 case LDB_OP_EXTENDED
:
456 ret
= ldb_parse_extended(ret
, attr
, value
);
470 <and> ::= '&' <filterlist>
471 <or> ::= '|' <filterlist>
472 <filterlist> ::= <filter> | <filter> <filterlist>
474 static struct ldb_parse_tree
*ldb_parse_filterlist(void *mem_ctx
, const char **s
)
476 struct ldb_parse_tree
*ret
, *next
;
477 enum ldb_parse_op op
;
492 while (isspace((unsigned char)*p
)) p
++;
494 ret
= talloc(mem_ctx
, struct ldb_parse_tree
);
501 ret
->u
.list
.num_elements
= 1;
502 ret
->u
.list
.elements
= talloc(ret
, struct ldb_parse_tree
*);
503 if (!ret
->u
.list
.elements
) {
509 ret
->u
.list
.elements
[0] = ldb_parse_filter(ret
->u
.list
.elements
, &p
);
510 if (!ret
->u
.list
.elements
[0]) {
515 while (isspace((unsigned char)*p
)) p
++;
517 while (*p
&& (next
= ldb_parse_filter(ret
->u
.list
.elements
, &p
))) {
518 struct ldb_parse_tree
**e
;
519 e
= talloc_realloc(ret
, ret
->u
.list
.elements
,
520 struct ldb_parse_tree
*,
521 ret
->u
.list
.num_elements
+ 1);
527 ret
->u
.list
.elements
= e
;
528 ret
->u
.list
.elements
[ret
->u
.list
.num_elements
] = next
;
529 ret
->u
.list
.num_elements
++;
530 while (isspace((unsigned char)*p
)) p
++;
540 <not> ::= '!' <filter>
542 static struct ldb_parse_tree
*ldb_parse_not(void *mem_ctx
, const char **s
)
544 struct ldb_parse_tree
*ret
;
552 ret
= talloc(mem_ctx
, struct ldb_parse_tree
);
558 ret
->operation
= LDB_OP_NOT
;
559 ret
->u
.isnot
.child
= ldb_parse_filter(ret
, &p
);
560 if (!ret
->u
.isnot
.child
) {
572 <filtercomp> ::= <and> | <or> | <not> | <simple>
574 static struct ldb_parse_tree
*ldb_parse_filtercomp(void *mem_ctx
, const char **s
)
576 struct ldb_parse_tree
*ret
;
579 while (isspace((unsigned char)*p
)) p
++;
583 ret
= ldb_parse_filterlist(mem_ctx
, &p
);
587 ret
= ldb_parse_filterlist(mem_ctx
, &p
);
591 ret
= ldb_parse_not(mem_ctx
, &p
);
599 ret
= ldb_parse_simple(mem_ctx
, &p
);
609 <filter> ::= '(' <filtercomp> ')'
611 static struct ldb_parse_tree
*ldb_parse_filter(void *mem_ctx
, const char **s
)
613 struct ldb_parse_tree
*ret
;
621 ret
= ldb_parse_filtercomp(mem_ctx
, &p
);
628 while (isspace((unsigned char)*p
)) {
639 main parser entry point. Takes a search string and returns a parse tree
641 expression ::= <simple> | <filter>
643 struct ldb_parse_tree
*ldb_parse_tree(void *mem_ctx
, const char *s
)
645 if (s
== NULL
|| *s
== 0) {
646 s
= "(|(objectClass=*)(distinguishedName=*))";
649 while (isspace((unsigned char)*s
)) s
++;
652 return ldb_parse_filter(mem_ctx
, &s
);
655 return ldb_parse_simple(mem_ctx
, &s
);
660 construct a ldap parse filter given a parse tree
662 char *ldb_filter_from_tree(void *mem_ctx
, struct ldb_parse_tree
*tree
)
671 switch (tree
->operation
) {
674 ret
= talloc_asprintf(mem_ctx
, "(%c", tree
->operation
==LDB_OP_AND
?'&':'|');
675 if (ret
== NULL
) return NULL
;
676 for (i
=0;i
<tree
->u
.list
.num_elements
;i
++) {
677 s
= ldb_filter_from_tree(mem_ctx
, tree
->u
.list
.elements
[i
]);
682 s2
= talloc_asprintf_append(ret
, "%s", s
);
690 s
= talloc_asprintf_append(ret
, ")");
697 s
= ldb_filter_from_tree(mem_ctx
, tree
->u
.isnot
.child
);
698 if (s
== NULL
) return NULL
;
700 ret
= talloc_asprintf(mem_ctx
, "(!%s)", s
);
703 case LDB_OP_EQUALITY
:
704 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
705 if (s
== NULL
) return NULL
;
706 ret
= talloc_asprintf(mem_ctx
, "(%s=%s)",
707 tree
->u
.equality
.attr
, s
);
710 case LDB_OP_SUBSTRING
:
711 ret
= talloc_asprintf(mem_ctx
, "(%s=%s", tree
->u
.substring
.attr
,
712 tree
->u
.substring
.start_with_wildcard
?"*":"");
713 if (ret
== NULL
) return NULL
;
714 for (i
= 0; tree
->u
.substring
.chunks
[i
]; i
++) {
715 s2
= ldb_binary_encode(mem_ctx
, *(tree
->u
.substring
.chunks
[i
]));
720 if (tree
->u
.substring
.chunks
[i
+1] ||
721 tree
->u
.substring
.end_with_wildcard
) {
722 s
= talloc_asprintf_append(ret
, "%s*", s2
);
724 s
= talloc_asprintf_append(ret
, "%s", s2
);
732 s
= talloc_asprintf_append(ret
, ")");
740 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
741 if (s
== NULL
) return NULL
;
742 ret
= talloc_asprintf(mem_ctx
, "(%s>=%s)",
743 tree
->u
.equality
.attr
, s
);
747 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
748 if (s
== NULL
) return NULL
;
749 ret
= talloc_asprintf(mem_ctx
, "(%s<=%s)",
750 tree
->u
.equality
.attr
, s
);
754 ret
= talloc_asprintf(mem_ctx
, "(%s=*)", tree
->u
.present
.attr
);
757 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
758 if (s
== NULL
) return NULL
;
759 ret
= talloc_asprintf(mem_ctx
, "(%s~=%s)",
760 tree
->u
.equality
.attr
, s
);
763 case LDB_OP_EXTENDED
:
764 s
= ldb_binary_encode(mem_ctx
, tree
->u
.extended
.value
);
765 if (s
== NULL
) return NULL
;
766 ret
= talloc_asprintf(mem_ctx
, "(%s%s%s%s:=%s)",
767 tree
->u
.extended
.attr
?tree
->u
.extended
.attr
:"",
768 tree
->u
.extended
.dnAttributes
?":dn":"",
769 tree
->u
.extended
.rule_id
?":":"",
770 tree
->u
.extended
.rule_id
?tree
->u
.extended
.rule_id
:"",
781 replace any occurances of an attribute name in the parse tree with a
784 void ldb_parse_tree_attr_replace(struct ldb_parse_tree
*tree
,
789 switch (tree
->operation
) {
792 for (i
=0;i
<tree
->u
.list
.num_elements
;i
++) {
793 ldb_parse_tree_attr_replace(tree
->u
.list
.elements
[i
],
798 ldb_parse_tree_attr_replace(tree
->u
.isnot
.child
, attr
, replace
);
800 case LDB_OP_EQUALITY
:
804 if (ldb_attr_cmp(tree
->u
.equality
.attr
, attr
) == 0) {
805 tree
->u
.equality
.attr
= replace
;
808 case LDB_OP_SUBSTRING
:
809 if (ldb_attr_cmp(tree
->u
.substring
.attr
, attr
) == 0) {
810 tree
->u
.substring
.attr
= replace
;
814 if (ldb_attr_cmp(tree
->u
.present
.attr
, attr
) == 0) {
815 tree
->u
.present
.attr
= replace
;
818 case LDB_OP_EXTENDED
:
819 if (tree
->u
.extended
.attr
&&
820 ldb_attr_cmp(tree
->u
.extended
.attr
, attr
) == 0) {
821 tree
->u
.extended
.attr
= replace
;