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;
97 encode a blob as a RFC2254 binary string, escaping any
98 non-printable or '\' characters
100 char *ldb_binary_encode(void *mem_ctx
, struct ldb_val val
)
104 int len
= val
.length
;
105 unsigned char *buf
= val
.data
;
107 for (i
=0;i
<val
.length
;i
++) {
108 if (!isprint(buf
[i
]) || strchr(" *()\\&|!\"", buf
[i
])) {
112 ret
= talloc_array(mem_ctx
, char, len
+1);
113 if (ret
== NULL
) return NULL
;
116 for (i
=0;i
<val
.length
;i
++) {
117 if (!isprint(buf
[i
]) || strchr(" *()\\&|!\"", buf
[i
])) {
118 snprintf(ret
+len
, 4, "\\%02X", buf
[i
]);
131 encode a string as a RFC2254 binary string, escaping any
132 non-printable or '\' characters. This routine is suitable for use
133 in escaping user data in ldap filters.
135 char *ldb_binary_encode_string(void *mem_ctx
, const char *string
)
138 val
.data
= discard_const_p(uint8_t, string
);
139 val
.length
= strlen(string
);
140 return ldb_binary_encode(mem_ctx
, val
);
143 /* find the first matching wildcard */
144 static char *ldb_parse_find_wildcard(char *value
)
147 value
= strpbrk(value
, "\\*");
148 if (value
== NULL
) return NULL
;
150 if (value
[0] == '\\') {
151 if (value
[1] == '\0') return NULL
;
156 if (value
[0] == '*') return value
;
162 /* return a NULL terminated list of binary strings representing the value
163 chunks separated by wildcards that makes the value portion of the filter
165 static struct ldb_val
**ldb_wildcard_decode(void *mem_ctx
, const char *string
)
167 struct ldb_val
**ret
= NULL
;
171 wc
= talloc_strdup(mem_ctx
, string
);
172 if (wc
== NULL
) return NULL
;
176 wc
= ldb_parse_find_wildcard(str
);
186 ret
= talloc_realloc(mem_ctx
, ret
, struct ldb_val
*, val
+ 2);
187 if (ret
== NULL
) return NULL
;
189 ret
[val
] = talloc(mem_ctx
, struct ldb_val
);
190 if (ret
[val
] == NULL
) return NULL
;
192 *(ret
[val
]) = ldb_binary_decode(mem_ctx
, str
);
193 if ((ret
[val
])->data
== NULL
) return NULL
;
205 static struct ldb_parse_tree
*ldb_parse_filter(void *mem_ctx
, const char **s
);
209 parse an extended match
217 the ':dn' part sets the dnAttributes boolean if present
218 the oid sets the rule_id string
221 static struct ldb_parse_tree
*ldb_parse_extended(struct ldb_parse_tree
*ret
,
222 char *attr
, char *value
)
226 ret
->operation
= LDB_OP_EXTENDED
;
227 ret
->u
.extended
.value
= ldb_binary_decode(ret
, value
);
228 if (ret
->u
.extended
.value
.data
== NULL
) goto failed
;
230 p1
= strchr(attr
, ':');
231 if (p1
== NULL
) goto failed
;
232 p2
= strchr(p1
+1, ':');
237 ret
->u
.extended
.attr
= attr
;
238 if (strcmp(p1
+1, "dn") == 0) {
239 ret
->u
.extended
.dnAttributes
= 1;
241 ret
->u
.extended
.rule_id
= talloc_strdup(ret
, p2
+1);
242 if (ret
->u
.extended
.rule_id
== NULL
) goto failed
;
244 ret
->u
.extended
.rule_id
= NULL
;
247 ret
->u
.extended
.dnAttributes
= 0;
248 ret
->u
.extended
.rule_id
= talloc_strdup(ret
, p1
+1);
249 if (ret
->u
.extended
.rule_id
== NULL
) goto failed
;
259 static enum ldb_parse_op
ldb_parse_filtertype(void *mem_ctx
, char **type
, char **value
, const char **s
)
261 enum ldb_parse_op filter
= 0;
262 char *name
, *val
, *k
;
266 /* retrieve attributetype name */
269 while ((isascii(*p
) && isalnum((unsigned char)*p
)) || (*p
== '-')) { /* attribute names can only be alphanums */
273 if (*p
== ':') { /* but extended searches have : and . chars too */
275 if (p
== NULL
) { /* malformed attribute name */
282 while (isspace((unsigned char)*p
)) p
++;
284 if (!strchr("=<>~:", *p
)) {
289 name
= (char *)talloc_memdup(mem_ctx
, t
, t1
- t
+ 1);
290 if (name
== NULL
) return 0;
293 /* retrieve filtertype */
296 filter
= LDB_OP_EQUALITY
;
297 } else if (*(p
+ 1) == '=') {
300 filter
= LDB_OP_LESS
;
304 filter
= LDB_OP_GREATER
;
308 filter
= LDB_OP_APPROX
;
312 filter
= LDB_OP_EXTENDED
;
323 while (isspace((unsigned char)*p
)) p
++;
328 while (*p
&& ((*p
!= ')') || ((*p
== ')') && (*(p
- 1) == '\\')))) p
++;
330 val
= (char *)talloc_memdup(mem_ctx
, t
, p
- t
+ 1);
339 /* remove trailing spaces from value */
340 while ((k
> val
) && (isspace((unsigned char)*(k
- 1)))) k
--;
350 <simple> ::= <attributetype> <filtertype> <attributevalue>
352 static struct ldb_parse_tree
*ldb_parse_simple(void *mem_ctx
, const char **s
)
355 struct ldb_parse_tree
*ret
;
356 enum ldb_parse_op filtertype
;
358 ret
= talloc(mem_ctx
, struct ldb_parse_tree
);
364 filtertype
= ldb_parse_filtertype(ret
, &attr
, &value
, s
);
370 switch (filtertype
) {
373 ret
->operation
= LDB_OP_PRESENT
;
374 ret
->u
.present
.attr
= attr
;
377 case LDB_OP_EQUALITY
:
379 if (strcmp(value
, "*") == 0) {
380 ret
->operation
= LDB_OP_PRESENT
;
381 ret
->u
.present
.attr
= attr
;
385 if (ldb_parse_find_wildcard(value
) != NULL
) {
386 ret
->operation
= LDB_OP_SUBSTRING
;
387 ret
->u
.substring
.attr
= attr
;
388 ret
->u
.substring
.start_with_wildcard
= 0;
389 ret
->u
.substring
.end_with_wildcard
= 0;
390 ret
->u
.substring
.chunks
= ldb_wildcard_decode(ret
, value
);
391 if (ret
->u
.substring
.chunks
== NULL
){
396 ret
->u
.substring
.start_with_wildcard
= 1;
397 if (value
[strlen(value
) - 1] == '*')
398 ret
->u
.substring
.end_with_wildcard
= 1;
404 ret
->operation
= LDB_OP_EQUALITY
;
405 ret
->u
.equality
.attr
= attr
;
406 ret
->u
.equality
.value
= ldb_binary_decode(ret
, value
);
407 if (ret
->u
.equality
.value
.data
== NULL
) {
415 ret
->operation
= LDB_OP_GREATER
;
416 ret
->u
.comparison
.attr
= attr
;
417 ret
->u
.comparison
.value
= ldb_binary_decode(ret
, value
);
418 if (ret
->u
.comparison
.value
.data
== NULL
) {
426 ret
->operation
= LDB_OP_LESS
;
427 ret
->u
.comparison
.attr
= attr
;
428 ret
->u
.comparison
.value
= ldb_binary_decode(ret
, value
);
429 if (ret
->u
.comparison
.value
.data
== NULL
) {
437 ret
->operation
= LDB_OP_APPROX
;
438 ret
->u
.comparison
.attr
= attr
;
439 ret
->u
.comparison
.value
= ldb_binary_decode(ret
, value
);
440 if (ret
->u
.comparison
.value
.data
== NULL
) {
447 case LDB_OP_EXTENDED
:
449 ret
= ldb_parse_extended(ret
, attr
, value
);
463 <and> ::= '&' <filterlist>
464 <or> ::= '|' <filterlist>
465 <filterlist> ::= <filter> | <filter> <filterlist>
467 static struct ldb_parse_tree
*ldb_parse_filterlist(void *mem_ctx
, const char **s
)
469 struct ldb_parse_tree
*ret
, *next
;
470 enum ldb_parse_op op
;
485 while (isspace((unsigned char)*p
)) p
++;
487 ret
= talloc(mem_ctx
, struct ldb_parse_tree
);
494 ret
->u
.list
.num_elements
= 1;
495 ret
->u
.list
.elements
= talloc(ret
, struct ldb_parse_tree
*);
496 if (!ret
->u
.list
.elements
) {
502 ret
->u
.list
.elements
[0] = ldb_parse_filter(ret
->u
.list
.elements
, &p
);
503 if (!ret
->u
.list
.elements
[0]) {
508 while (isspace((unsigned char)*p
)) p
++;
510 while (*p
&& (next
= ldb_parse_filter(ret
->u
.list
.elements
, &p
))) {
511 struct ldb_parse_tree
**e
;
512 e
= talloc_realloc(ret
, ret
->u
.list
.elements
,
513 struct ldb_parse_tree
*,
514 ret
->u
.list
.num_elements
+ 1);
520 ret
->u
.list
.elements
= e
;
521 ret
->u
.list
.elements
[ret
->u
.list
.num_elements
] = next
;
522 ret
->u
.list
.num_elements
++;
523 while (isspace((unsigned char)*p
)) p
++;
533 <not> ::= '!' <filter>
535 static struct ldb_parse_tree
*ldb_parse_not(void *mem_ctx
, const char **s
)
537 struct ldb_parse_tree
*ret
;
545 ret
= talloc(mem_ctx
, struct ldb_parse_tree
);
551 ret
->operation
= LDB_OP_NOT
;
552 ret
->u
.isnot
.child
= ldb_parse_filter(ret
, &p
);
553 if (!ret
->u
.isnot
.child
) {
565 <filtercomp> ::= <and> | <or> | <not> | <simple>
567 static struct ldb_parse_tree
*ldb_parse_filtercomp(void *mem_ctx
, const char **s
)
569 struct ldb_parse_tree
*ret
;
572 while (isspace((unsigned char)*p
)) p
++;
576 ret
= ldb_parse_filterlist(mem_ctx
, &p
);
580 ret
= ldb_parse_filterlist(mem_ctx
, &p
);
584 ret
= ldb_parse_not(mem_ctx
, &p
);
592 ret
= ldb_parse_simple(mem_ctx
, &p
);
602 <filter> ::= '(' <filtercomp> ')'
604 static struct ldb_parse_tree
*ldb_parse_filter(void *mem_ctx
, const char **s
)
606 struct ldb_parse_tree
*ret
;
614 ret
= ldb_parse_filtercomp(mem_ctx
, &p
);
621 while (isspace((unsigned char)*p
)) {
632 main parser entry point. Takes a search string and returns a parse tree
634 expression ::= <simple> | <filter>
636 struct ldb_parse_tree
*ldb_parse_tree(void *mem_ctx
, const char *s
)
638 if (s
== NULL
|| *s
== 0) {
639 s
= "(|(objectClass=*)(distinguishedName=*))";
642 while (isspace((unsigned char)*s
)) s
++;
645 return ldb_parse_filter(mem_ctx
, &s
);
648 return ldb_parse_simple(mem_ctx
, &s
);
653 construct a ldap parse filter given a parse tree
655 char *ldb_filter_from_tree(void *mem_ctx
, struct ldb_parse_tree
*tree
)
664 switch (tree
->operation
) {
667 ret
= talloc_asprintf(mem_ctx
, "(%c", tree
->operation
==LDB_OP_AND
?'&':'|');
668 if (ret
== NULL
) return NULL
;
669 for (i
=0;i
<tree
->u
.list
.num_elements
;i
++) {
670 s
= ldb_filter_from_tree(mem_ctx
, tree
->u
.list
.elements
[i
]);
675 s2
= talloc_asprintf_append(ret
, "%s", s
);
683 s
= talloc_asprintf_append(ret
, ")");
690 s
= ldb_filter_from_tree(mem_ctx
, tree
->u
.isnot
.child
);
691 if (s
== NULL
) return NULL
;
693 ret
= talloc_asprintf(mem_ctx
, "(!%s)", s
);
696 case LDB_OP_EQUALITY
:
697 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
698 if (s
== NULL
) return NULL
;
699 ret
= talloc_asprintf(mem_ctx
, "(%s=%s)",
700 tree
->u
.equality
.attr
, s
);
703 case LDB_OP_SUBSTRING
:
704 ret
= talloc_asprintf(mem_ctx
, "(%s=%s", tree
->u
.substring
.attr
,
705 tree
->u
.substring
.start_with_wildcard
?"*":"");
706 if (ret
== NULL
) return NULL
;
707 for (i
= 0; tree
->u
.substring
.chunks
[i
]; i
++) {
708 s2
= ldb_binary_encode(mem_ctx
, *(tree
->u
.substring
.chunks
[i
]));
713 if (tree
->u
.substring
.chunks
[i
+1] ||
714 tree
->u
.substring
.end_with_wildcard
) {
715 s
= talloc_asprintf_append(ret
, "%s*", s2
);
717 s
= talloc_asprintf_append(ret
, "%s", s2
);
725 s
= talloc_asprintf_append(ret
, ")");
733 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
734 if (s
== NULL
) return NULL
;
735 ret
= talloc_asprintf(mem_ctx
, "(%s>=%s)",
736 tree
->u
.equality
.attr
, s
);
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 ret
= talloc_asprintf(mem_ctx
, "(%s=*)", tree
->u
.present
.attr
);
750 s
= ldb_binary_encode(mem_ctx
, tree
->u
.equality
.value
);
751 if (s
== NULL
) return NULL
;
752 ret
= talloc_asprintf(mem_ctx
, "(%s~=%s)",
753 tree
->u
.equality
.attr
, s
);
756 case LDB_OP_EXTENDED
:
757 s
= ldb_binary_encode(mem_ctx
, tree
->u
.extended
.value
);
758 if (s
== NULL
) return NULL
;
759 ret
= talloc_asprintf(mem_ctx
, "(%s%s%s%s:=%s)",
760 tree
->u
.extended
.attr
?tree
->u
.extended
.attr
:"",
761 tree
->u
.extended
.dnAttributes
?":dn":"",
762 tree
->u
.extended
.rule_id
?":":"",
763 tree
->u
.extended
.rule_id
?tree
->u
.extended
.rule_id
:"",
774 replace any occurances of an attribute name in the parse tree with a
777 void ldb_parse_tree_attr_replace(struct ldb_parse_tree
*tree
,
782 switch (tree
->operation
) {
785 for (i
=0;i
<tree
->u
.list
.num_elements
;i
++) {
786 ldb_parse_tree_attr_replace(tree
->u
.list
.elements
[i
],
791 ldb_parse_tree_attr_replace(tree
->u
.isnot
.child
, attr
, replace
);
793 case LDB_OP_EQUALITY
:
797 if (ldb_attr_cmp(tree
->u
.equality
.attr
, attr
) == 0) {
798 tree
->u
.equality
.attr
= replace
;
801 case LDB_OP_SUBSTRING
:
802 if (ldb_attr_cmp(tree
->u
.substring
.attr
, attr
) == 0) {
803 tree
->u
.substring
.attr
= replace
;
807 if (ldb_attr_cmp(tree
->u
.present
.attr
, attr
) == 0) {
808 tree
->u
.present
.attr
= replace
;
811 case LDB_OP_EXTENDED
:
812 if (tree
->u
.extended
.attr
&&
813 ldb_attr_cmp(tree
->u
.extended
.attr
, attr
) == 0) {
814 tree
->u
.extended
.attr
= replace
;