schema_init.c 105 KB
Newer Older
1
2
3
/* schema_init.c - init builtin schema */
/* $OpenLDAP$ */
/*
Kurt Zeilenga's avatar
Kurt Zeilenga committed
4
 * Copyright 1998-2000 The OpenLDAP Foundation, All Rights Reserved.
5
6
7
8
9
10
 * COPYING RESTRICTIONS APPLY, see COPYRIGHT file
 */

#include "portable.h"

#include <stdio.h>
Kurt Zeilenga's avatar
Kurt Zeilenga committed
11
#include <limits.h>
12
13

#include <ac/ctype.h>
14
#include <ac/errno.h>
15
16
17
18
19
#include <ac/string.h>
#include <ac/socket.h>

#include "slap.h"
#include "ldap_pvt.h"
20

21
22
#include "ldap_utf8.h"

23
24
25
26
27
28
#include "lutil_hash.h"
#define HASH_BYTES				LUTIL_HASH_BYTES
#define HASH_CONTEXT			lutil_HASH_CTX
#define HASH_Init(c)			lutil_HASHInit(c)
#define HASH_Update(c,buf,len)	lutil_HASHUpdate(c,buf,len)
#define HASH_Final(d,c)			lutil_HASHFinal(d,c)
29

30
/* recycled validatation routines */
31
#define berValidate						blobValidate
32
33

/* unimplemented pretters */
34
#define integerPretty					NULL
35
36

/* recycled matching routines */
37
#define bitStringMatch					octetStringMatch
38
39
40
#define numericStringMatch				caseIgnoreIA5Match
#define objectIdentifierMatch			caseIgnoreIA5Match
#define telephoneNumberMatch			caseIgnoreIA5Match
41
#define telephoneNumberSubstringsMatch	caseIgnoreIA5SubstringsMatch
42
43
#define generalizedTimeMatch			caseIgnoreIA5Match
#define generalizedTimeOrderingMatch	caseIgnoreIA5Match
44
#define uniqueMemberMatch				dnMatch
45

46
47
/* approx matching rules */
#define directoryStringApproxMatchOID	"1.3.6.1.4.1.4203.666.4.4"
Gary Williams's avatar
Gary Williams committed
48
49
50
#define directoryStringApproxMatch	approxMatch
#define directoryStringApproxIndexer	approxIndexer
#define directoryStringApproxFilter	approxFilter
51
#define IA5StringApproxMatchOID			"1.3.6.1.4.1.4203.666.4.5"
Gary Williams's avatar
Gary Williams committed
52
#define IA5StringApproxMatch			approxMatch
53
#define IA5StringApproxIndexer			approxIndexer
Gary Williams's avatar
Gary Williams committed
54
#define IA5StringApproxFilter			approxFilter
55

56
57
58
59
/* orderring matching rules */
#define caseIgnoreOrderingMatch			caseIgnoreMatch
#define caseExactOrderingMatch			caseExactMatch

60
/* unimplemented matching routines */
61
62
63
64
65
66
67
#define caseIgnoreListMatch				NULL
#define caseIgnoreListSubstringsMatch	NULL
#define protocolInformationMatch		NULL
#define integerFirstComponentMatch		NULL

#define OpenLDAPaciMatch				NULL
#define authPasswordMatch				NULL
68
69

/* recycled indexing/filtering routines */
70
71
#define dnIndexer				caseExactIgnoreIndexer
#define dnFilter				caseExactIgnoreFilter
72
73
#define bitStringFilter			octetStringFilter
#define bitStringIndexer		octetStringIndexer
74

75
76
77
78
79
#define telephoneNumberIndexer			caseIgnoreIA5Indexer
#define telephoneNumberFilter			caseIgnoreIA5Filter
#define telephoneNumberSubstringsIndexer	caseIgnoreIA5SubstringsIndexer
#define telephoneNumberSubstringsFilter		caseIgnoreIA5SubstringsFilter

80
81
82
83
/* must match OIDs below */
#define caseExactMatchOID			"2.5.13.5"
#define caseExactSubstringsMatchOID		"2.5.13.7"

84
85
86
87
static char *strcasechr( const char *str, int c )
{
	char *lower = strchr( str, TOLOWER(c) );
	char *upper = strchr( str, TOUPPER(c) );
88

89
90
91
92
93
94
95
96
	if( lower && upper ) {
		return lower < upper ? lower : upper;
	} else if ( lower ) {
		return lower;
	} else {
		return upper;
	}
}
97

98
99
100
static int
octetStringMatch(
	int *matchp,
101
	slap_mask_t flags,
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
{
	int match = value->bv_len - ((struct berval *) assertedValue)->bv_len;

	if( match == 0 ) {
		match = memcmp( value->bv_val,
			((struct berval *) assertedValue)->bv_val,
			value->bv_len );
	}

	*matchp = match;
	return LDAP_SUCCESS;
}

/* Index generation function */
120
static int octetStringIndexer(
121
122
	slap_mask_t use,
	slap_mask_t flags,
123
124
125
126
127
128
129
130
131
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	struct berval **values,
	struct berval ***keysp )
{
	int i;
	size_t slen, mlen;
	struct berval **keys;
132
	HASH_CONTEXT   HASHcontext;
Gary Williams's avatar
Gary Williams committed
133
	unsigned char	HASHdigest[HASH_BYTES];
134
	struct berval digest;
135
136
	digest.bv_val = HASHdigest;
	digest.bv_len = sizeof(HASHdigest);
137
138
139
140
141

	for( i=0; values[i] != NULL; i++ ) {
		/* just count them */
	}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
142
143
144
	/* we should have at least one value at this point */
	assert( i > 0 );

145
146
147
148
149
150
	keys = ch_malloc( sizeof( struct berval * ) * (i+1) );

	slen = strlen( syntax->ssyn_oid );
	mlen = strlen( mr->smr_oid );

	for( i=0; values[i] != NULL; i++ ) {
151
		HASH_Init( &HASHcontext );
152
		if( prefix != NULL && prefix->bv_len > 0 ) {
153
			HASH_Update( &HASHcontext,
154
155
				prefix->bv_val, prefix->bv_len );
		}
156
		HASH_Update( &HASHcontext,
157
			syntax->ssyn_oid, slen );
158
		HASH_Update( &HASHcontext,
159
			mr->smr_oid, mlen );
160
		HASH_Update( &HASHcontext,
161
			values[i]->bv_val, values[i]->bv_len );
162
		HASH_Final( HASHdigest, &HASHcontext );
163
164
165
166
167
168
169
170
171
172
173
174

		keys[i] = ber_bvdup( &digest );
	}

	keys[i] = NULL;

	*keysp = keys;

	return LDAP_SUCCESS;
}

/* Index generation function */
175
static int octetStringFilter(
176
177
	slap_mask_t use,
	slap_mask_t flags,
178
179
180
181
182
183
184
185
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	void * assertValue,
	struct berval ***keysp )
{
	size_t slen, mlen;
	struct berval **keys;
186
	HASH_CONTEXT   HASHcontext;
Gary Williams's avatar
Gary Williams committed
187
	unsigned char	HASHdigest[HASH_BYTES];
188
189
	struct berval *value = (struct berval *) assertValue;
	struct berval digest;
190
191
	digest.bv_val = HASHdigest;
	digest.bv_len = sizeof(HASHdigest);
192
193
194
195
196
197

	slen = strlen( syntax->ssyn_oid );
	mlen = strlen( mr->smr_oid );

	keys = ch_malloc( sizeof( struct berval * ) * 2 );

198
	HASH_Init( &HASHcontext );
199
	if( prefix != NULL && prefix->bv_len > 0 ) {
200
		HASH_Update( &HASHcontext,
201
202
			prefix->bv_val, prefix->bv_len );
	}
203
	HASH_Update( &HASHcontext,
204
		syntax->ssyn_oid, slen );
205
	HASH_Update( &HASHcontext,
206
		mr->smr_oid, mlen );
207
	HASH_Update( &HASHcontext,
208
		value->bv_val, value->bv_len );
209
	HASH_Final( HASHdigest, &HASHcontext );
210
211
212
213
214
215
216
217

	keys[0] = ber_bvdup( &digest );
	keys[1] = NULL;

	*keysp = keys;

	return LDAP_SUCCESS;
}
218

219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
static int
nameUIDValidate(
	Syntax *syntax,
	struct berval *in )
{
	int rc;
	struct berval *dn;

	if( in->bv_len == 0 ) return LDAP_SUCCESS;

	dn = ber_bvdup( in );

	if( dn->bv_val[dn->bv_len-1] == '\'' ) {
		/* assume presence of optional UID */
		ber_len_t i;

		for(i=dn->bv_len-2; i>2; i--) {
			if( dn->bv_val[i] != '0' &&	dn->bv_val[i] != '1' ) {
				break;
			}
		}
Stig Venaas's avatar
Stig Venaas committed
240
241
242
243
		if( dn->bv_val[i] != '\'' ||
		    dn->bv_val[i-1] != 'B' ||
		    dn->bv_val[i-2] != '#' ) {
			ber_bvfree( dn );
244
245
246
247
248
249
250
			return LDAP_INVALID_SYNTAX;
		}

		/* trim the UID to allow use of dn_validate */
		dn->bv_val[i-2] = '\0';
	}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
251
	/* FIXME: should use dnValidate */
252
253
254
255
256
257
258
259
260
261
262
263
264
265
	rc = dn_validate( dn->bv_val ) == NULL
		? LDAP_INVALID_SYNTAX : LDAP_SUCCESS;

	ber_bvfree( dn );
	return rc;
}

static int
nameUIDNormalize(
	Syntax *syntax,
	struct berval *val,
	struct berval **normalized )
{
	struct berval *out = ber_bvdup( val );
Howard Chu's avatar
Howard Chu committed
266
	int rc;
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282

	if( out->bv_len != 0 ) {
		char *dn;
		ber_len_t dnlen;
		char *uid = NULL;
		ber_len_t uidlen = 0;

		if( out->bv_val[out->bv_len-1] == '\'' ) {
			/* assume presence of optional UID */
			uid = strrchr( out->bv_val, '#' );

			if( uid == NULL ) {
				ber_bvfree( out );
				return LDAP_INVALID_SYNTAX;
			}

Howard Chu's avatar
Howard Chu committed
283
			uidlen = out->bv_len - (uid - out->bv_val);
284
285
			/* temporarily trim the UID */
			*uid = '\0';
Howard Chu's avatar
Howard Chu committed
286
			out->bv_len -= uidlen;
287
288
289
		}

#ifdef USE_DN_NORMALIZE
Howard Chu's avatar
Howard Chu committed
290
		rc = dnNormalize( NULL, out, normalized );
291
#else
Howard Chu's avatar
Howard Chu committed
292
		rc = dnPretty( NULL, out, normalized );
293
294
#endif

Howard Chu's avatar
Howard Chu committed
295
		if( rc != LDAP_SUCCESS ) {
296
297
298
299
			ber_bvfree( out );
			return LDAP_INVALID_SYNTAX;
		}

Howard Chu's avatar
Howard Chu committed
300
		dnlen = (*normalized)->bv_len;
301
302

		if( uidlen ) {
Howard Chu's avatar
Howard Chu committed
303
304
305
306
			struct berval *b2 = ch_malloc(sizeof(struct berval));
			b2->bv_val = ch_malloc(dnlen + uidlen + 1);
			SAFEMEMCPY( b2->bv_val, (*normalized)->bv_val, dnlen );

307
308
309
			/* restore the separator */
			*uid = '#';
			/* shift the UID */
Howard Chu's avatar
Howard Chu committed
310
311
312
313
314
			SAFEMEMCPY( (*normalized)->bv_val+dnlen, uid, uidlen );
			b2->bv_len = dnlen + uidlen;
			(*normalized)->bv_val[dnlen+uidlen] = '\0';
			ber_bvfree(*normalized);
			*normalized = b2;
315
		}
Howard Chu's avatar
Howard Chu committed
316
		ber_bvfree( out );
317
318
319
320
321
	}

	return LDAP_SUCCESS;
}

322
323
324
325
326
327
328
329
330
static int
inValidate(
	Syntax *syntax,
	struct berval *in )
{
	/* any value allowed */
	return LDAP_OTHER;
}

331
static int
332
blobValidate(
333
334
335
336
	Syntax *syntax,
	struct berval *in )
{
	/* any value allowed */
337
	return LDAP_SUCCESS;
338
339
}

340
341
342
343
344
345
346
347
348
349
350
351
352
static int
bitStringValidate(
	Syntax *syntax,
	struct berval *in )
{
	ber_len_t i;

	/* very unforgiving validation, requires no normalization
	 * before simplistic matching
	 */
	if( in->bv_len < 3 ) {
		return LDAP_INVALID_SYNTAX;
	}
353

354
355
356
357
358
359
360
361
362
363
	/*
	 * rfc 2252 section 6.3 Bit String
	 * bitstring = "'" *binary-digit "'"
	 * binary-digit = "0" / "1"
	 * example: '0101111101'B
	 */
	
	if( in->bv_val[0] != '\'' ||
		in->bv_val[in->bv_len-2] != '\'' ||
		in->bv_val[in->bv_len-1] != 'B' )
364
365
366
367
	{
		return LDAP_INVALID_SYNTAX;
	}

368
	for( i=in->bv_len-3; i>0; i-- ) {
369
370
371
372
373
374
375
376
		if( in->bv_val[i] != '0' && in->bv_val[i] != '1' ) {
			return LDAP_INVALID_SYNTAX;
		}
	}

	return LDAP_SUCCESS;
}

377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
static int
bitStringNormalize(
	Syntax *syntax,
	struct berval *val,
	struct berval **normalized )
{
	/*
     * A normalized bitString is has no extaneous (leading) zero bits.
	 * That is, '00010'B is normalized to '10'B
	 * However, as a special case, '0'B requires no normalization.
     */
	struct berval *newval;
	char *p;

	/* start at the first bit */
	p = &val->bv_val[1];

	/* Find the first non-zero bit */
	while ( *p == '0' ) p++;

	newval = (struct berval *) ch_malloc( sizeof(struct berval) );

	if( *p == '\'' ) {
		/* no non-zero bits */
		newval->bv_val = ch_strdup("\'0\'B");
		newval->bv_len = sizeof("\'0\'B") - 1;
		goto done;
	}

	newval->bv_val = ch_malloc( val->bv_len + 1 );

	newval->bv_val[0] = '\'';
	newval->bv_len = 1;

	for( ; *p != '\0'; p++ ) {
		newval->bv_val[newval->bv_len++] = *p;
	}

	newval->bv_val[newval->bv_len] = '\0';

done:
	*normalized = newval;
	return LDAP_SUCCESS;
}

422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
/*
 * Handling boolean syntax and matching is quite rigid.
 * A more flexible approach would be to allow a variety
 * of strings to be normalized and prettied into TRUE
 * and FALSE.
 */
static int
booleanValidate(
	Syntax *syntax,
	struct berval *in )
{
	/* very unforgiving validation, requires no normalization
	 * before simplistic matching
	 */

	if( in->bv_len == 4 ) {
		if( !memcmp( in->bv_val, "TRUE", 4 ) ) {
			return LDAP_SUCCESS;
		}
	} else if( in->bv_len == 5 ) {
		if( !memcmp( in->bv_val, "FALSE", 5 ) ) {
			return LDAP_SUCCESS;
		}
	}

	return LDAP_INVALID_SYNTAX;
}

static int
booleanMatch(
	int *matchp,
453
	slap_mask_t flags,
454
455
456
457
458
459
460
461
462
463
464
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
{
	/* simplistic matching allowed by rigid validation */
	struct berval *asserted = (struct berval *) assertedValue;
	*matchp = value->bv_len != asserted->bv_len;
	return LDAP_SUCCESS;
}

465
466
467
468
469
470
471
472
473
static int
UTF8StringValidate(
	Syntax *syntax,
	struct berval *in )
{
	ber_len_t count;
	int len;
	unsigned char *u = in->bv_val;

474
475
	if( !in->bv_len ) return LDAP_INVALID_SYNTAX;

476
	for( count = in->bv_len; count > 0; count-=len, u+=len ) {
477
478
479
480
		/* get the length indicated by the first byte */
		len = LDAP_UTF8_CHARLEN( u );

		/* should not be zero */
481
		if( len == 0 ) return LDAP_INVALID_SYNTAX;
482
483
484

		/* make sure len corresponds with the offset
			to the next character */
485
		if( LDAP_UTF8_OFFSET( u ) != len ) return LDAP_INVALID_SYNTAX;
486
487
	}

488
	if( count != 0 ) return LDAP_INVALID_SYNTAX;
489

490
	return LDAP_SUCCESS;
491
492
493
494
495
496
497
498
499
}

static int
UTF8StringNormalize(
	Syntax *syntax,
	struct berval *val,
	struct berval **normalized )
{
	struct berval *newval;
500
	char *p, *q, *s;
501

502
	newval = ch_malloc( sizeof( struct berval ) );
503

504
	p = val->bv_val;
505

506
507
508
	/* Ignore initial whitespace */
	while ( ldap_utf8_isspace( p ) ) {
		LDAP_UTF8_INCR( p );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
509
	}
510

511
512
	if( *p == '\0' ) {
		ch_free( newval );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
513
514
		return LDAP_INVALID_SYNTAX;
	}
515

516
517
518
	newval->bv_val = ch_strdup( p );
	p = q = newval->bv_val;
	s = NULL;
519

520
521
	while ( *p ) {
		int len;
522

523
524
525
526
527
		if ( ldap_utf8_isspace( p ) ) {
			len = LDAP_UTF8_COPY(q,p);
			s=q;
			p+=len;
			q+=len;
528

529
530
531
532
			/* Ignore the extra whitespace */
			while ( ldap_utf8_isspace( p ) ) {
				LDAP_UTF8_INCR( p );
			}
Kurt Zeilenga's avatar
Kurt Zeilenga committed
533
		} else {
534
535
536
537
			len = LDAP_UTF8_COPY(q,p);
			s=NULL;
			p+=len;
			q+=len;
Kurt Zeilenga's avatar
Kurt Zeilenga committed
538
		}
539
540
541
542
	}

	assert( *newval->bv_val );
	assert( newval->bv_val < p );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
543
	assert( q <= p );
544

545
546
547
548
549
550
551
552
553
554
555
	/* cannot start with a space */
	assert( !ldap_utf8_isspace(newval->bv_val) );

	/*
	 * If the string ended in space, backup the pointer one
	 * position.  One is enough because the above loop collapsed
	 * all whitespace to a single space.
	 */

	if ( s != NULL ) {
		q = s;
Kurt Zeilenga's avatar
Kurt Zeilenga committed
556
	}
557

558
559
560
561
562
563
564
565
	/* cannot end with a space */
	assert( !ldap_utf8_isspace( LDAP_UTF8_PREV(q) ) );

	/* null terminate */
	*q = '\0';

	newval->bv_len = q - newval->bv_val;
	*normalized = newval;
566

567
	return LDAP_SUCCESS;
568
569
}

570
571
/* Returns Unicode cannonically normalized copy of a substring assertion
 * Skipping attribute description */
572
static SubstringsAssertion *
573
574
575
576
577
578
579
580
581
582
583
584
585
UTF8SubstringsassertionNormalize(
	SubstringsAssertion *sa,
	char casefold )
{
	SubstringsAssertion *nsa;
	int i;

	nsa = (SubstringsAssertion *)ch_calloc( 1, sizeof(SubstringsAssertion) );
	if( nsa == NULL ) {
		return NULL;
	}

	if( sa->sa_initial != NULL ) {
586
		nsa->sa_initial = ber_bvstr( UTF8normalize( sa->sa_initial, casefold ) );
587
588
589
590
591
592
593
594
595
596
597
		if( nsa->sa_initial == NULL ) {
			goto err;
		}
	}

	if( sa->sa_any != NULL ) {
		for( i=0; sa->sa_any[i] != NULL; i++ ) {
			/* empty */
		}
		nsa->sa_any = (struct berval **)ch_malloc( (i + 1) * sizeof(struct berval *) );
		for( i=0; sa->sa_any[i] != NULL; i++ ) {
598
			nsa->sa_any[i] = ber_bvstr( UTF8normalize( sa->sa_any[i], casefold ) );
599
600
601
602
603
604
605
606
			if( nsa->sa_any[i] == NULL ) {
				goto err;
			}
		}
		nsa->sa_any[i] = NULL;
	}

	if( sa->sa_final != NULL ) {
607
		nsa->sa_final = ber_bvstr( UTF8normalize( sa->sa_final, casefold ) );
608
609
610
611
612
613
614
615
		if( nsa->sa_final == NULL ) {
			goto err;
		}
	}

	return nsa;

err:
616
	ber_bvfree( nsa->sa_final );
617
	ber_bvecfree( nsa->sa_any );
618
	ber_bvfree( nsa->sa_initial );
619
620
621
622
	ch_free( nsa );
	return NULL;
}

623
/* Strip characters with the 8th bit set */
624
static char *
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
strip8bitChars(
	char *in )      
{
	char *p = in, *q;
  
	if( in == NULL ) {
		return NULL;
	}
	while( *p ) {
		if( *p & 0x80 ) {
			q = p;
			while( *++q & 0x80 ) {
				/* empty */
			}
			p = memmove(p, q, strlen(q) + 1);
		} else {
			p++;
		}
	}
	return in;
}

647
#ifndef SLAPD_APPROX_OLDSINGLESTRING
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665

#if defined(SLAPD_APPROX_INITIALS)
#define SLAPD_APPROX_DELIMITER "._ "
#define SLAPD_APPROX_WORDLEN 2
#else
#define SLAPD_APPROX_DELIMITER " "
#define SLAPD_APPROX_WORDLEN 1
#endif

static int
approxMatch(
	int *matchp,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
{
666
	char *val, *nval, *assertv, **values, **words, *c;
667
	int i, count, len, nextchunk=0, nextavail=0;
668
	size_t avlen;
669

670
	/* Yes, this is necessary */
671
	nval = UTF8normalize( value, UTF8_NOCASEFOLD );
672
	if( nval == NULL ) {
673
674
675
		*matchp = 1;
		return LDAP_SUCCESS;
	}
676
	strip8bitChars( nval );
677
678

	/* Yes, this is necessary */
679
	assertv = UTF8normalize( ((struct berval *)assertedValue),
680
681
				 UTF8_NOCASEFOLD );
	if( assertv == NULL ) {
682
		ch_free( nval );
683
684
685
686
687
		*matchp = 1;
		return LDAP_SUCCESS;
	}
	strip8bitChars( assertv );
	avlen = strlen( assertv );
688
689

	/* Isolate how many words there are */
690
	for( c=nval,count=1; *c; c++ ) {
691
692
693
694
695
696
697
698
699
		c = strpbrk( c, SLAPD_APPROX_DELIMITER );
		if ( c == NULL ) break;
		*c = '\0';
		count++;
	}

	/* Get a phonetic copy of each word */
	words = (char **)ch_malloc( count * sizeof(char *) );
	values = (char **)ch_malloc( count * sizeof(char *) );
700
	for( c=nval,i=0;  i<count;  i++,c+=strlen(c)+1 ) {
701
702
703
704
		words[i] = c;
		values[i] = phonetic(c);
	}

705
	/* Work through the asserted value's words, to see if at least some
706
707
	   of the words are there, in the same order. */
	len = 0;
708
	while ( (size_t) nextchunk < avlen ) {
709
		len = strcspn( assertv + nextchunk, SLAPD_APPROX_DELIMITER);
710
711
712
713
		if( len == 0 ) {
			nextchunk++;
			continue;
		}
714
#if defined(SLAPD_APPROX_INITIALS)
715
		else if( len == 1 ) {
716
717
			/* Single letter words need to at least match one word's initial */
			for( i=nextavail; i<count; i++ )
718
719
				if( !strncasecmp( assertv+nextchunk, words[i], 1 )) {
					nextavail=i+1;
720
					break;
721
				}
722
723
		}
#endif
724
		else {
725
726
727
728
729
730
731
732
733
734
735
			/* Isolate the next word in the asserted value and phonetic it */
			assertv[nextchunk+len] = '\0';
			val = phonetic( assertv + nextchunk );

			/* See if this phonetic chunk is in the remaining words of *value */
			for( i=nextavail; i<count; i++ ){
				if( !strcmp( val, values[i] ) ){
					nextavail = i+1;
					break;
				}
			}
736
			ch_free( val );
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
		}

		/* This chunk in the asserted value was NOT within the *value. */
		if( i >= count ) {
			nextavail=-1;
			break;
		}

		/* Go on to the next word in the asserted value */
		nextchunk += len+1;
	}

	/* If some of the words were seen, call it a match */
	if( nextavail > 0 ) {
		*matchp = 0;
	}
	else {
		*matchp = 1;
	}

	/* Cleanup allocs */
758
	free( assertv );
759
760
761
762
763
	for( i=0; i<count; i++ ) {
		ch_free( values[i] );
	}
	ch_free( values );
	ch_free( words );
764
	ch_free( nval );
765
766
767
768

	return LDAP_SUCCESS;
}

769
static int 
770
771
772
773
774
775
776
777
778
779
780
781
782
783
approxIndexer(
	slap_mask_t use,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	struct berval **values,
	struct berval ***keysp )
{
	char *val, *c;
	int i,j, len, wordcount, keycount=0;
	struct berval **newkeys, **keys=NULL;

	for( j=0; values[j] != NULL; j++ ) {
784
		/* Yes, this is necessary */
785
		val = UTF8normalize( values[j], UTF8_NOCASEFOLD );
786
787
		strip8bitChars( val );

788
		/* Isolate how many words there are. There will be a key for each */
Gary Williams's avatar
Gary Williams committed
789
		for( wordcount=0,c=val;	 *c;  c++) {
790
791
792
793
794
795
796
797
798
			len = strcspn(c, SLAPD_APPROX_DELIMITER);
			if( len >= SLAPD_APPROX_WORDLEN ) wordcount++;
			c+= len;
			if (*c == '\0') break;
			*c = '\0';
		}

		/* Allocate/increase storage to account for new keys */
		newkeys = (struct berval **)ch_malloc( (keycount + wordcount + 1) 
Kurt Zeilenga's avatar
Kurt Zeilenga committed
799
			* sizeof(struct berval *) );
800
801
802
803
804
		memcpy( newkeys, keys, keycount * sizeof(struct berval *) );
		if( keys ) ch_free( keys );
		keys = newkeys;

		/* Get a phonetic copy of each word */
Gary Williams's avatar
Gary Williams committed
805
		for( c=val,i=0;	 i<wordcount;  c+=len+1	 ) {
806
807
808
809
810
811
812
813
814
			len = strlen( c );
			if( len < SLAPD_APPROX_WORDLEN ) continue;
			keys[keycount] = (struct berval *)ch_malloc( sizeof(struct berval) );
			keys[keycount]->bv_val = phonetic( c );
			keys[keycount]->bv_len = strlen( keys[keycount]->bv_val );
			keycount++;
			i++;
		}

815
		free( val );
816
817
818
819
820
821
822
	}
	keys[keycount] = NULL;
	*keysp = keys;

	return LDAP_SUCCESS;
}

823
static int 
824
825
826
827
828
829
830
831
832
833
834
835
836
approxFilter(
	slap_mask_t use,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	void * assertValue,
	struct berval ***keysp )
{
	char *val, *c;
	int i, count, len;
	struct berval **keys;

837
	/* Yes, this is necessary */
838
	val = UTF8normalize( ((struct berval *)assertValue),
839
840
841
842
843
844
845
846
847
			     UTF8_NOCASEFOLD );
	if( val == NULL ) {
		keys = (struct berval **)ch_malloc( sizeof(struct berval *) );
		keys[0] = NULL;
		*keysp = keys;
		return LDAP_SUCCESS;
	}
	strip8bitChars( val );

848
849
850
851
852
853
854
855
856
857
858
859
860
	/* Isolate how many words there are. There will be a key for each */
	for( count=0,c=val;  *c;  c++) {
		len = strcspn(c, SLAPD_APPROX_DELIMITER);
		if( len >= SLAPD_APPROX_WORDLEN ) count++;
		c+= len;
		if (*c == '\0') break;
		*c = '\0';
	}

	/* Allocate storage for new keys */
	keys = (struct berval **)ch_malloc( (count + 1) * sizeof(struct berval *) );

	/* Get a phonetic copy of each word */
Gary Williams's avatar
Gary Williams committed
861
	for( c=val,i=0;	 i<count; c+=len+1 ) {
862
863
		len = strlen(c);
		if( len < SLAPD_APPROX_WORDLEN ) continue;
864
		keys[i] = ber_bvstr( phonetic( c ) );
865
866
867
		i++;
	}

868
	free( val );
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889

	keys[count] = NULL;
	*keysp = keys;

	return LDAP_SUCCESS;
}


#else
/* No other form of Approximate Matching is defined */

static int
approxMatch(
	int *matchp,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
{
	char *vapprox, *avapprox;
890
	char *s, *t;
891

892
	/* Yes, this is necessary */
893
	s = UTF8normalize( value, UTF8_NOCASEFOLD );
894
895
896
897
898
899
	if( s == NULL ) {
		*matchp = 1;
		return LDAP_SUCCESS;
	}

	/* Yes, this is necessary */
900
	t = UTF8normalize( ((struct berval *)assertedValue),
901
902
903
904
905
906
907
908
909
910
911
912
			   UTF8_NOCASEFOLD );
	if( t == NULL ) {
		free( s );
		*matchp = -1;
		return LDAP_SUCCESS;
	}

	vapprox = phonetic( strip8bitChars( s ) );
	avapprox = phonetic( strip8bitChars( t ) );

	free( s );
	free( t );
913
914
915
916
917
918
919
920
921

	*matchp = strcmp( vapprox, avapprox );

	ch_free( vapprox );
	ch_free( avapprox );

	return LDAP_SUCCESS;
}

922
static int 
923
924
925
926
927
928
929
930
931
932
933
approxIndexer(
	slap_mask_t use,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	struct berval **values,
	struct berval ***keysp )
{
	int i;
	struct berval **keys;
934
	char *s;
935
936

	for( i=0; values[i] != NULL; i++ ) {
Kurt Zeilenga's avatar
Kurt Zeilenga committed
937
		/* empty - just count them */
938
	}
Kurt Zeilenga's avatar
Kurt Zeilenga committed
939
940

	/* we should have at least one value at this point */
941
942
943
944
945
946
	assert( i > 0 );

	keys = (struct berval **)ch_malloc( sizeof( struct berval * ) * (i+1) );

	/* Copy each value and run it through phonetic() */
	for( i=0; values[i] != NULL; i++ ) {
947
		/* Yes, this is necessary */
948
		s = UTF8normalize( values[i], UTF8_NOCASEFOLD );
949
950
951
952

		/* strip 8-bit chars and run through phonetic() */
		keys[i] = ber_bvstr( phonetic( strip8bitChars( s ) ) );
		free( s );
953
954
955
956
957
958
959
960
	}
	keys[i] = NULL;

	*keysp = keys;
	return LDAP_SUCCESS;
}


961
static int 
962
963
964
965
966
967
968
969
970
971
approxFilter(
	slap_mask_t use,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	void * assertValue,
	struct berval ***keysp )
{
	struct berval **keys;
972
	char *s;
973
974
975

	keys = (struct berval **)ch_malloc( sizeof( struct berval * ) * 2 );

976
	/* Yes, this is necessary */
977
	s = UTF8normalize( ((struct berval *)assertValue),
978
979
980
981
982
983
984
985
986
			     UTF8_NOCASEFOLD );
	if( s == NULL ) {
		keys[0] = NULL;
	} else {
		/* strip 8-bit chars and run through phonetic() */
		keys[0] = ber_bvstr( phonetic( strip8bitChars( s ) ) );
		free( s );
		keys[1] = NULL;
	}
987
988
989
990
991
992
993

	*keysp = keys;
	return LDAP_SUCCESS;
}
#endif


994
static int
995
caseExactMatch(
996
	int *matchp,
997
	slap_mask_t flags,
998
999
1000
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
1001
	void *assertedValue )
1002
{
1003
1004
1005
	*matchp = UTF8normcmp( value->bv_val,
		((struct berval *) assertedValue)->bv_val,
		UTF8_NOCASEFOLD );
1006
	return LDAP_SUCCESS;
1007
1008
}

1009
static int
1010
caseExactIgnoreSubstringsMatch(
1011
	int *matchp,
1012
	slap_mask_t flags,
1013
1014
1015
1016
1017
1018
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
{
	int match = 0;
Pierangelo Masarati's avatar
Pierangelo Masarati committed
1019
	SubstringsAssertion *sub = NULL;
1020
	struct berval left;
1021
1022
	int i;
	ber_len_t inlen=0;
1023
1024
1025
1026
	char *nav, casefold;

	casefold = strcmp( mr->smr_oid, caseExactSubstringsMatchOID )
		? UTF8_CASEFOLD : UTF8_NOCASEFOLD;
1027

1028
	nav = UTF8normalize( value, casefold );
1029
1030
1031
1032
1033
1034
	if( nav == NULL ) {
		match = 1;
		goto done;
	}
	left.bv_val = nav;
	left.bv_len = strlen( nav );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1035

1036
	sub = UTF8SubstringsassertionNormalize( assertedValue, casefold );
1037
1038
1039
1040
	if( sub == NULL ) {
		match = -1;
		goto done;
	}
1041

1042
	/* Add up asserted input length */
1043
1044
1045
1046
	if( sub->sa_initial ) {
		inlen += sub->sa_initial->bv_len;
	}
	if( sub->sa_any ) {
1047
1048
		for(i=0; sub->sa_any[i] != NULL; i++) {
			inlen += sub->sa_any[i]->bv_len;
1049
1050
1051
1052
1053
1054
1055
		}
	}
	if( sub->sa_final ) {
		inlen += sub->sa_final->bv_len;
	}

	if( sub->sa_initial ) {
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1056
1057
1058
1059
1060
		if( inlen > left.bv_len ) {
			match = 1;
			goto done;
		}

1061
1062
		match = strncmp( sub->sa_initial->bv_val, left.bv_val,
			sub->sa_initial->bv_len );
1063
1064
1065
1066
1067
1068
1069

		if( match != 0 ) {
			goto done;
		}

		left.bv_val += sub->sa_initial->bv_len;
		left.bv_len -= sub->sa_initial->bv_len;
1070
		inlen -= sub->sa_initial->bv_len;
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1071
	}
1072

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1073
	if( sub->sa_final ) {
1074
1075
1076
1077
		if( inlen > left.bv_len ) {
			match = 1;
			goto done;
		}
1078

1079
1080
1081
		match = strncmp( sub->sa_final->bv_val,
			&left.bv_val[left.bv_len - sub->sa_final->bv_len],
			sub->sa_final->bv_len );
1082
1083
1084
1085
1086
1087

		if( match != 0 ) {
			goto done;
		}

		left.bv_len -= sub->sa_final->bv_len;
1088
		inlen -= sub->sa_final->bv_len;
1089
1090
1091
	}

	if( sub->sa_any ) {
1092
1093
1094
1095
1096
		for(i=0; sub->sa_any[i]; i++) {
			ber_len_t idx;
			char *p;

retry:
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1097
1098
1099
1100
1101
1102
			if( inlen > left.bv_len ) {
				/* not enough length */
				match = 1;
				goto done;
			}

1103
1104
1105
1106
			if( sub->sa_any[i]->bv_len == 0 ) {
				continue;
			}

1107
			p = strchr( left.bv_val, *sub->sa_any[i]->bv_val );
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118

			if( p == NULL ) {
				match = 1;
				goto done;
			}

			idx = p - left.bv_val;
			assert( idx < left.bv_len );

			if( idx >= left.bv_len ) {
				/* this shouldn't happen */
Stig Venaas's avatar
Stig Venaas committed
1119
1120
1121
1122
1123
				free( nav );
				ch_free( sub->sa_final );
				ber_bvecfree( sub->sa_any );
				ch_free( sub->sa_initial );
				ch_free( sub );
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
				return LDAP_OTHER;
			}

			left.bv_val = p;
			left.bv_len -= idx;

			if( sub->sa_any[i]->bv_len > left.bv_len ) {
				/* not enough left */
				match = 1;
				goto done;
			}

1136
1137
1138
			match = strncmp( left.bv_val,
				sub->sa_any[i]->bv_val,
				sub->sa_any[i]->bv_len );
1139
1140

			if( match != 0 ) {
1141
1142
				left.bv_val++;
				left.bv_len--;
1143
1144
1145
1146
1147
				goto retry;
			}

			left.bv_val += sub->sa_any[i]->bv_len;
			left.bv_len -= sub->sa_any[i]->bv_len;
1148
			inlen -= sub->sa_any[i]->bv_len;
1149
		}
1150
1151
1152
	}

done:
1153
1154
	free( nav );
	if( sub != NULL ) {
1155
		ber_bvfree( sub->sa_final );
1156
		ber_bvecfree( sub->sa_any );
1157
		ber_bvfree( sub->sa_initial );
1158
1159
		ch_free( sub );
	}
1160
1161
1162
1163
	*matchp = match;
	return LDAP_SUCCESS;
}

1164
/* Index generation function */
1165
static int caseExactIgnoreIndexer(
1166
1167
	slap_mask_t use,
	slap_mask_t flags,
1168
1169
1170
1171
1172
1173
1174
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	struct berval **values,
	struct berval ***keysp )
{
	int i;
1175
	char casefold;
1176
1177
	size_t slen, mlen;
	struct berval **keys;
1178
	HASH_CONTEXT   HASHcontext;
Gary Williams's avatar
Gary Williams committed
1179
	unsigned char	HASHdigest[HASH_BYTES];
1180
	struct berval digest;
1181
1182
	digest.bv_val = HASHdigest;
	digest.bv_len = sizeof(HASHdigest);
1183
1184

	for( i=0; values[i] != NULL; i++ ) {
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1185
		/* empty - just count them */
1186
1187
	}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1188
1189
1190
	/* we should have at least one value at this point */
	assert( i > 0 );

1191
1192
1193
1194
1195
	keys = ch_malloc( sizeof( struct berval * ) * (i+1) );

	slen = strlen( syntax->ssyn_oid );
	mlen = strlen( mr->smr_oid );

1196
1197
1198
	casefold = strcmp( mr->smr_oid, caseExactMatchOID )
		? UTF8_CASEFOLD : UTF8_NOCASEFOLD;

1199
	for( i=0; values[i] != NULL; i++ ) {
1200
		struct berval *value;
1201
		value = ber_bvstr( UTF8normalize( values[i],
1202
			casefold ) );
1203

1204
		HASH_Init( &HASHcontext );
1205
		if( prefix != NULL && prefix->bv_len > 0 ) {
1206
			HASH_Update( &HASHcontext,
1207
1208
				prefix->bv_val, prefix->bv_len );
		}
1209
		HASH_Update( &HASHcontext,
1210
			syntax->ssyn_oid, slen );
1211
		HASH_Update( &HASHcontext,
1212
			mr->smr_oid, mlen );
1213
		HASH_Update( &HASHcontext,
1214
			value->bv_val, value->bv_len );
1215
		HASH_Final( HASHdigest, &HASHcontext );
1216

1217
1218
		ber_bvfree( value );

1219
1220
1221
1222
1223
1224
1225
1226
1227
		keys[i] = ber_bvdup( &digest );
	}

	keys[i] = NULL;
	*keysp = keys;
	return LDAP_SUCCESS;
}

/* Index generation function */
1228
static int caseExactIgnoreFilter(
1229
1230
	slap_mask_t use,
	slap_mask_t flags,
1231
1232
1233
1234
1235
1236
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	void * assertValue,
	struct berval ***keysp )
{
1237
	char casefold;
1238
1239
	size_t slen, mlen;
	struct berval **keys;
1240
	HASH_CONTEXT   HASHcontext;
Gary Williams's avatar
Gary Williams committed
1241
	unsigned char	HASHdigest[HASH_BYTES];
1242
1243
	struct berval *value;
	struct berval digest;
1244
1245
	digest.bv_val = HASHdigest;
	digest.bv_len = sizeof(HASHdigest);
1246
1247
1248
1249

	slen = strlen( syntax->ssyn_oid );
	mlen = strlen( mr->smr_oid );

1250
1251
1252
	casefold = strcmp( mr->smr_oid, caseExactMatchOID )
		? UTF8_CASEFOLD : UTF8_NOCASEFOLD;

1253
	value = ber_bvstr( UTF8normalize( ((struct berval *) assertValue),
1254
		casefold ) );
1255
1256
1257
1258
1259
1260
	/* This usually happens if filter contains bad UTF8 */
	if( value == NULL ) {
		keys = ch_malloc( sizeof( struct berval * ) );
		keys[0] = NULL;
		return LDAP_SUCCESS;
	}