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
43 #include "ldb_private.h"
44 #include "system/locale.h"
46 static int ldb_parse_hex2char(const char *x
)
48 if (isxdigit(x
[0]) && isxdigit(x
[1])) {
49 const char h1
= x
[0], h2
= x
[1];
52 if (h1
>= 'a') c
= h1
- (int)'a' + 10;
53 else if (h1
>= 'A') c
= h1
- (int)'A' + 10;
54 else if (h1
>= '0') c
= h1
- (int)'0';
56 if (h2
>= 'a') c
+= h2
- (int)'a' + 10;
57 else if (h2
>= 'A') c
+= h2
- (int)'A' + 10;
58 else if (h2
>= '0') c
+= h2
- (int)'0';
67 a filter is defined by:
68 <filter> ::= '(' <filtercomp> ')'
69 <filtercomp> ::= <and> | <or> | <not> | <simple>
70 <and> ::= '&' <filterlist>
71 <or> ::= '|' <filterlist>
72 <not> ::= '!' <filter>
73 <filterlist> ::= <filter> | <filter> <filterlist>
74 <simple> ::= <attributetype> <filtertype> <attributevalue>
75 <filtertype> ::= '=' | '~=' | '<=' | '>='
79 decode a RFC2254 binary string representation of a buffer.
82 struct ldb_val
ldb_binary_decode(TALLOC_CTX
*mem_ctx
, const char *str
)
86 size_t slen
= str
?strlen(str
):0;
88 ret
.data
= (uint8_t *)talloc_size(mem_ctx
, slen
+1);
90 if (ret
.data
== NULL
) return ret
;
92 for (i
=j
=0;i
<slen
;i
++) {
96 c
= ldb_parse_hex2char(&str
[i
+1]);
98 talloc_free(ret
.data
);
99 memset(&ret
, 0, sizeof(ret
));
102 ((uint8_t *)ret
.data
)[j
++] = c
;
105 ((uint8_t *)ret
.data
)[j
++] = str
[i
];
109 ((uint8_t *)ret
.data
)[j
] = 0;
116 encode a blob as a RFC2254 binary string, escaping any
117 non-printable or '\' characters
119 char *ldb_binary_encode(TALLOC_CTX
*mem_ctx
, struct ldb_val val
)
123 size_t len
= val
.length
;
124 unsigned char *buf
= val
.data
;
126 for (i
=0;i
<val
.length
;i
++) {
127 if (!isprint(buf
[i
]) || strchr(" *()\\&|!\"", buf
[i
])) {
131 ret
= talloc_array(mem_ctx
, char, len
+1);
132 if (ret
== NULL
) return NULL
;
135 for (i
=0;i
<val
.length
;i
++) {
136 if (!isprint(buf
[i
]) || strchr(" *()\\&|!\"", buf
[i
])) {
137 snprintf(ret
+len
, 4, "\\%02X", buf
[i
]);
150 encode a string as a RFC2254 binary string, escaping any
151 non-printable or '\' characters. This routine is suitable for use
152 in escaping user data in ldap filters.
154 char *ldb_binary_encode_string(TALLOC_CTX
*mem_ctx
, const char *string
)
157 if (string
== NULL
) {
160 val
.data
= discard_const_p(uint8_t, string
);
161 val
.length
= strlen(string
);
162 return ldb_binary_encode(mem_ctx
, val
);
165 /* find the first matching wildcard */
166 static char *ldb_parse_find_wildcard(char *value
)
169 value
= strpbrk(value
, "\\*");
170 if (value
== NULL
) return NULL
;
172 if (value
[0] == '\\') {
173 if (value
[1] == '\0') return NULL
;
178 if (value
[0] == '*') return value
;
184 /* return a NULL terminated list of binary strings representing the value
185 chunks separated by wildcards that makes the value portion of the filter
187 static struct ldb_val
**ldb_wildcard_decode(TALLOC_CTX
*mem_ctx
, const char *string
)
189 struct ldb_val
**ret
= NULL
;
190 unsigned int val
= 0;
193 wc
= talloc_strdup(mem_ctx
, string
);
194 if (wc
== NULL
) return NULL
;
198 wc
= ldb_parse_find_wildcard(str
);
208 ret
= talloc_realloc(mem_ctx
, ret
, struct ldb_val
*, val
+ 2);
209 if (ret
== NULL
) return NULL
;
211 ret
[val
] = talloc(mem_ctx
, struct ldb_val
);
212 if (ret
[val
] == NULL
) return NULL
;
214 *(ret
[val
]) = ldb_binary_decode(mem_ctx
, str
);
215 if ((ret
[val
])->data
== NULL
) return NULL
;
227 static struct ldb_parse_tree
*ldb_parse_filter(TALLOC_CTX
*mem_ctx
, const char **s
);
231 parse an extended match
239 the ':dn' part sets the dnAttributes boolean if present
240 the oid sets the rule_id string
243 static struct ldb_parse_tree
*ldb_parse_extended(struct ldb_parse_tree
*ret
,
244 char *attr
, char *value
)
248 ret
->operation
= LDB_OP_EXTENDED
;
249 ret
->u
.extended
.value
= ldb_binary_decode(ret
, value
);
250 if (ret
->u
.extended
.value
.data
== NULL
) goto failed
;
252 p1
= strchr(attr
, ':');
253 if (p1
== NULL
) goto failed
;
254 p2
= strchr(p1
+1, ':');
259 ret
->u
.extended
.attr
= attr
;
260 if (strcmp(p1
+1, "dn") == 0) {
261 ret
->u
.extended
.dnAttributes
= 1;
263 ret
->u
.extended
.rule_id
= talloc_strdup(ret
, p2
+1);
264 if (ret
->u
.extended
.rule_id
== NULL
) goto failed
;
266 ret
->u
.extended
.rule_id
= NULL
;
269 ret
->u
.extended
.dnAttributes
= 0;
270 ret
->u
.extended
.rule_id
= talloc_strdup(ret
, p1
+1);
271 if (ret
->u
.extended
.rule_id
== NULL
) goto failed
;
281 static enum ldb_parse_op
ldb_parse_filtertype(TALLOC_CTX
*mem_ctx
, char **type
, char **value
, const char **s
)
283 enum ldb_parse_op filter
= 0;
284 char *name
, *val
, *k
;
288 /* retrieve attributetype name */
291 if (*p
== '@') { /* for internal attributes the first char can be @ */
295 while ((isascii(*p
) && isalnum((unsigned char)*p
)) || (*p
== '-') || (*p
== '.')) {
296 /* attribute names can only be alphanums */
300 if (*p
== ':') { /* but extended searches have : and . chars too */
302 if (p
== NULL
) { /* malformed attribute name */
309 while (isspace((unsigned char)*p
)) p
++;
311 if (!strchr("=<>~:", *p
)) {
316 name
= (char *)talloc_memdup(mem_ctx
, t
, t1
- t
+ 1);
317 if (name
== NULL
) return 0;
320 /* retrieve filtertype */
323 filter
= LDB_OP_EQUALITY
;
324 } else if (*(p
+ 1) == '=') {
327 filter
= LDB_OP_LESS
;
331 filter
= LDB_OP_GREATER
;
335 filter
= LDB_OP_APPROX
;
339 filter
= LDB_OP_EXTENDED
;
350 while (isspace((unsigned char)*p
)) p
++;
355 while (*p
&& ((*p
!= ')') || ((*p
== ')') && (*(p
- 1) == '\\')))) p
++;
357 val
= (char *)talloc_memdup(mem_ctx
, t
, p
- t
+ 1);
366 /* remove trailing spaces from value */
367 while ((k
> val
) && (isspace((unsigned char)*(k
- 1)))) k
--;
377 <simple> ::= <attributetype> <filtertype> <attributevalue>
379 static struct ldb_parse_tree
*ldb_parse_simple(TALLOC_CTX
*mem_ctx
, const char **s
)
382 struct ldb_parse_tree
*ret
;
383 enum ldb_parse_op filtertype
;
385 ret
= talloc(mem_ctx
, struct ldb_parse_tree
);
391 filtertype
= ldb_parse_filtertype(ret
, &attr
, &value
, s
);
397 switch (filtertype
) {
400 ret
->operation
= LDB_OP_PRESENT
;
401 ret
->u
.present
.attr
= attr
;
404 case LDB_OP_EQUALITY
:
406 if (strcmp(value
, "*") == 0) {
407 ret
->operation
= LDB_OP_PRESENT
;
408 ret
->u
.present
.attr
= attr
;
412 if (ldb_parse_find_wildcard(value
) != NULL
) {
413 ret
->operation
= LDB_OP_SUBSTRING
;
414 ret
->u
.substring
.attr
= attr
;
415 ret
->u
.substring
.start_with_wildcard
= 0;
416 ret
->u
.substring
.end_with_wildcard
= 0;
417 ret
->u
.substring
.chunks
= ldb_wildcard_decode(ret
, value
);
418 if (ret
->u
.substring
.chunks
== NULL
){
423 ret
->u
.substring
.start_with_wildcard
= 1;
424 if (value
[strlen(value
) - 1] == '*')
425 ret
->u
.substring
.end_with_wildcard
= 1;
431 ret
->operation
= LDB_OP_EQUALITY
;
432 ret
->u
.equality
.attr
= attr
;
433 ret
->u
.equality
.value
= ldb_binary_decode(ret
, value
);
434 if (ret
->u
.equality
.value
.data
== NULL
) {
442 ret
->operation
= LDB_OP_GREATER
;
443 ret
->u
.comparison
.attr
= attr
;
444 ret
->u
.comparison
.value
= ldb_binary_decode(ret
, value
);
445 if (ret
->u
.comparison
.value
.data
== NULL
) {
453 ret
->operation
= LDB_OP_LESS
;
454 ret
->u
.comparison
.attr
= attr
;
455 ret
->u
.comparison
.value
= ldb_binary_decode(ret
, value
);
456 if (ret
->u
.comparison
.value
.data
== NULL
) {
464 ret
->operation
= LDB_OP_APPROX
;
465 ret
->u
.comparison
.attr
= attr
;
466 ret
->u
.comparison
.value
= ldb_binary_decode(ret
, value
);
467 if (ret
->u
.comparison
.value
.data
== NULL
) {
474 case LDB_OP_EXTENDED
:
476 ret
= ldb_parse_extended(ret
, attr
, value
);
490 <and> ::= '&' <filterlist>
491 <or> ::= '|' <filterlist>
492 <filterlist> ::= <filter> | <filter> <filterlist>
494 static struct ldb_parse_tree
*ldb_parse_filterlist(TALLOC_CTX
*mem_ctx
, const char **s
)
496 struct ldb_parse_tree
*ret
, *next
;
497 enum ldb_parse_op op
;
512 while (isspace((unsigned char)*p
)) p
++;
514 ret
= talloc(mem_ctx
, struct ldb_parse_tree
);
521 ret
->u
.list
.num_elements
= 1;
522 ret
->u
.list
.elements
= talloc(ret
, struct ldb_parse_tree
*);
523 if (!ret
->u
.list
.elements
) {
529 ret
->u
.list
.elements
[0] = ldb_parse_filter(ret
->u
.list
.elements
, &p
);
530 if (!ret
->u
.list
.elements
[0]) {
535 while (isspace((unsigned char)*p
)) p
++;
538 struct ldb_parse_tree
**e
;
543 next
= ldb_parse_filter(ret
->u
.list
.elements
, &p
);
545 /* an invalid filter element */
549 e
= talloc_realloc(ret
, ret
->u
.list
.elements
,
550 struct ldb_parse_tree
*,
551 ret
->u
.list
.num_elements
+ 1);
557 ret
->u
.list
.elements
= e
;
558 ret
->u
.list
.elements
[ret
->u
.list
.num_elements
] = next
;
559 ret
->u
.list
.num_elements
++;
560 while (isspace((unsigned char)*p
)) p
++;
570 <not> ::= '!' <filter>
572 static struct ldb_parse_tree
*ldb_parse_not(TALLOC_CTX
*mem_ctx
, const char **s
)
574 struct ldb_parse_tree
*ret
;
582 ret
= talloc(mem_ctx
, struct ldb_parse_tree
);
588 ret
->operation
= LDB_OP_NOT
;
589 ret
->u
.isnot
.child
= ldb_parse_filter(ret
, &p
);
590 if (!ret
->u
.isnot
.child
) {
602 <filtercomp> ::= <and> | <or> | <not> | <simple>
604 static struct ldb_parse_tree
*ldb_parse_filtercomp(TALLOC_CTX
*mem_ctx
, const char **s
)
606 struct ldb_parse_tree
*ret
;
609 while (isspace((unsigned char)*p
)) p
++;
613 ret
= ldb_parse_filterlist(mem_ctx
, &p
);
617 ret
= ldb_parse_filterlist(mem_ctx
, &p
);
621 ret
= ldb_parse_not(mem_ctx
, &p
);
629 ret
= ldb_parse_simple(mem_ctx
, &p
);
639 <filter> ::= '(' <filtercomp> ')'
641 static struct ldb_parse_tree
*ldb_parse_filter(TALLOC_CTX
*mem_ctx
, const char **s
)
643 struct ldb_parse_tree
*ret
;
651 ret
= ldb_parse_filtercomp(mem_ctx
, &p
);
658 while (isspace((unsigned char)*p
)) {
669 main parser entry point. Takes a search string and returns a parse tree
671 expression ::= <simple> | <filter>
673 struct ldb_parse_tree
*ldb_parse_tree(TALLOC_CTX
*mem_ctx
, const char *s
)
675 if (s
== NULL
|| *s
== 0) {
676 s
= "(|(objectClass=*)(distinguishedName=*))";
679 while (isspace((unsigned char)*s
)) s
++;
682 return ldb_parse_filter(mem_ctx
, &s
);
685 return ldb_parse_simple(mem_ctx
, &s
);
690 construct a ldap parse filter given a parse tree
692 char *ldb_filter_from_tree(TALLOC_CTX
*mem_ctx
, const struct ldb_parse_tree
*tree
)
701 switch (tree
->operation
) {
704 ret
= talloc_asprintf(mem_ctx
, "(%c", tree
->operation
==LDB_OP_AND
?'&':'|');
705 if (ret
== NULL
) return NULL
;
706 for (i
=0;i
<tree
->u
.list
.num_elements
;i
++) {
707 s
= ldb_filter_from_tree(mem_ctx
, tree
->u
.list
.elements
[i
]);
712 s2
= talloc_asprintf_append(ret
, "%s", s
);
720 s
= talloc_asprintf_append(ret
, ")");
727 s
= ldb_filter_from_tree(mem_ctx
, tree
->u
.isnot
.child
);
728 if (s
== NULL
) return NULL
;
730 ret
= talloc_asprintf(mem_ctx
, "(!%s)", s
);
733 case LDB_OP_EQUALITY
:
734 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
735 if (s
== NULL
) return NULL
;
736 ret
= talloc_asprintf(mem_ctx
, "(%s=%s)",
737 tree
->u
.equality
.attr
, s
);
740 case LDB_OP_SUBSTRING
:
741 ret
= talloc_asprintf(mem_ctx
, "(%s=%s", tree
->u
.substring
.attr
,
742 tree
->u
.substring
.start_with_wildcard
?"*":"");
743 if (ret
== NULL
) return NULL
;
744 for (i
= 0; tree
->u
.substring
.chunks
[i
]; i
++) {
745 s2
= ldb_binary_encode(mem_ctx
, *(tree
->u
.substring
.chunks
[i
]));
750 if (tree
->u
.substring
.chunks
[i
+1] ||
751 tree
->u
.substring
.end_with_wildcard
) {
752 s
= talloc_asprintf_append(ret
, "%s*", s2
);
754 s
= talloc_asprintf_append(ret
, "%s", s2
);
762 s
= talloc_asprintf_append(ret
, ")");
770 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
771 if (s
== NULL
) return NULL
;
772 ret
= talloc_asprintf(mem_ctx
, "(%s>=%s)",
773 tree
->u
.equality
.attr
, s
);
777 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
778 if (s
== NULL
) return NULL
;
779 ret
= talloc_asprintf(mem_ctx
, "(%s<=%s)",
780 tree
->u
.equality
.attr
, s
);
784 ret
= talloc_asprintf(mem_ctx
, "(%s=*)", tree
->u
.present
.attr
);
787 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
788 if (s
== NULL
) return NULL
;
789 ret
= talloc_asprintf(mem_ctx
, "(%s~=%s)",
790 tree
->u
.equality
.attr
, s
);
793 case LDB_OP_EXTENDED
:
794 s
= ldb_binary_encode(mem_ctx
, tree
->u
.extended
.value
);
795 if (s
== NULL
) return NULL
;
796 ret
= talloc_asprintf(mem_ctx
, "(%s%s%s%s:=%s)",
797 tree
->u
.extended
.attr
?tree
->u
.extended
.attr
:"",
798 tree
->u
.extended
.dnAttributes
?":dn":"",
799 tree
->u
.extended
.rule_id
?":":"",
800 tree
->u
.extended
.rule_id
?tree
->u
.extended
.rule_id
:"",
811 walk a parse tree, calling the provided callback on each node
813 int ldb_parse_tree_walk(struct ldb_parse_tree
*tree
,
814 int (*callback
)(struct ldb_parse_tree
*tree
, void *),
815 void *private_context
)
820 ret
= callback(tree
, private_context
);
821 if (ret
!= LDB_SUCCESS
) {
825 switch (tree
->operation
) {
828 for (i
=0;i
<tree
->u
.list
.num_elements
;i
++) {
829 ret
= ldb_parse_tree_walk(tree
->u
.list
.elements
[i
], callback
, private_context
);
830 if (ret
!= LDB_SUCCESS
) {
836 ret
= ldb_parse_tree_walk(tree
->u
.isnot
.child
, callback
, private_context
);
837 if (ret
!= LDB_SUCCESS
) {
841 case LDB_OP_EQUALITY
:
845 case LDB_OP_SUBSTRING
:
847 case LDB_OP_EXTENDED
:
853 struct parse_tree_attr_replace_ctx
{
859 callback for ldb_parse_tree_attr_replace()
861 static int parse_tree_attr_replace(struct ldb_parse_tree
*tree
, void *private_context
)
863 struct parse_tree_attr_replace_ctx
*ctx
= private_context
;
864 switch (tree
->operation
) {
865 case LDB_OP_EQUALITY
:
869 if (ldb_attr_cmp(tree
->u
.equality
.attr
, ctx
->attr
) == 0) {
870 tree
->u
.equality
.attr
= ctx
->replace
;
873 case LDB_OP_SUBSTRING
:
874 if (ldb_attr_cmp(tree
->u
.substring
.attr
, ctx
->attr
) == 0) {
875 tree
->u
.substring
.attr
= ctx
->replace
;
879 if (ldb_attr_cmp(tree
->u
.present
.attr
, ctx
->attr
) == 0) {
880 tree
->u
.present
.attr
= ctx
->replace
;
883 case LDB_OP_EXTENDED
:
884 if (tree
->u
.extended
.attr
&&
885 ldb_attr_cmp(tree
->u
.extended
.attr
, ctx
->attr
) == 0) {
886 tree
->u
.extended
.attr
= ctx
->replace
;
896 replace any occurrences of an attribute name in the parse tree with a
899 void ldb_parse_tree_attr_replace(struct ldb_parse_tree
*tree
,
903 struct parse_tree_attr_replace_ctx ctx
;
906 ctx
.replace
= replace
;
908 ldb_parse_tree_walk(tree
, parse_tree_attr_replace
, &ctx
);
912 shallow copy a tree - copying only the elements array so that the caller
913 can safely add new elements without changing the message
915 struct ldb_parse_tree
*ldb_parse_tree_copy_shallow(TALLOC_CTX
*mem_ctx
,
916 const struct ldb_parse_tree
*ot
)
919 struct ldb_parse_tree
*nt
;
921 nt
= talloc(mem_ctx
, struct ldb_parse_tree
);
928 switch (ot
->operation
) {
931 nt
->u
.list
.elements
= talloc_array(nt
, struct ldb_parse_tree
*,
932 ot
->u
.list
.num_elements
);
933 if (!nt
->u
.list
.elements
) {
938 for (i
=0;i
<ot
->u
.list
.num_elements
;i
++) {
939 nt
->u
.list
.elements
[i
] =
940 ldb_parse_tree_copy_shallow(nt
->u
.list
.elements
,
941 ot
->u
.list
.elements
[i
]);
942 if (!nt
->u
.list
.elements
[i
]) {
949 nt
->u
.isnot
.child
= ldb_parse_tree_copy_shallow(nt
,
951 if (!nt
->u
.isnot
.child
) {
956 case LDB_OP_EQUALITY
:
960 case LDB_OP_SUBSTRING
:
962 case LDB_OP_EXTENDED
: