schema_init.c 71.7 KB
Newer Older
1
2
3
/* schema_init.c - init builtin schema */
/* $OpenLDAP$ */
/*
Kurt Zeilenga's avatar
Kurt Zeilenga committed
4
 * Copyright 1998-2003 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"
Pierangelo Masarati's avatar
Pierangelo Masarati committed
20
#include "lber_pvt.h"
21

22
23
#include "ldap_utf8.h"

24
25
26
27
28
29
#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)
30

Pierangelo Masarati's avatar
Pierangelo Masarati committed
31
32
#define	OpenLDAPaciMatch			NULL

33
34
/* approx matching rules */
#define directoryStringApproxMatchOID	"1.3.6.1.4.1.4203.666.4.4"
35
#define directoryStringApproxMatch		approxMatch
Gary Williams's avatar
Gary Williams committed
36
#define directoryStringApproxIndexer	approxIndexer
37
#define directoryStringApproxFilter		approxFilter
38
#define IA5StringApproxMatchOID			"1.3.6.1.4.1.4203.666.4.5"
Gary Williams's avatar
Gary Williams committed
39
#define IA5StringApproxMatch			approxMatch
40
#define IA5StringApproxIndexer			approxIndexer
Gary Williams's avatar
Gary Williams committed
41
#define IA5StringApproxFilter			approxFilter
42

Kurt Zeilenga's avatar
Kurt Zeilenga committed
43
44
45
46
static int
inValidate(
	Syntax *syntax,
	struct berval *in )
47
{
Kurt Zeilenga's avatar
Kurt Zeilenga committed
48
49
50
	/* no value allowed */
	return LDAP_INVALID_SYNTAX;
}
51

Kurt Zeilenga's avatar
Kurt Zeilenga committed
52
53
54
55
56
57
58
static int
blobValidate(
	Syntax *syntax,
	struct berval *in )
{
	/* any value allowed */
	return LDAP_SUCCESS;
59
}
60

Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
61
62
#define berValidate blobValidate

63
64
65
static int
octetStringMatch(
	int *matchp,
66
	slap_mask_t flags,
67
68
69
70
71
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
{
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
72
73
	struct berval *asserted = (struct berval *) assertedValue;
	int match = value->bv_len - asserted->bv_len;
74
75

	if( match == 0 ) {
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
76
		match = memcmp( value->bv_val, asserted->bv_val, value->bv_len );
77
78
79
80
81
82
	}

	*matchp = match;
	return LDAP_SUCCESS;
}

83
84
85
86
87
88
89
90
91
static int
octetStringOrderingMatch(
	int *matchp,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
{
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
92
	struct berval *asserted = (struct berval *) assertedValue;
93
	ber_len_t v_len  = value->bv_len;
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
94
	ber_len_t av_len = asserted->bv_len;
95

Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
96
	int match = memcmp( value->bv_val, asserted->bv_val,
97
		(v_len < av_len ? v_len : av_len) );
98
99
100

	if( match == 0 ) match = v_len - av_len;

101
102
103
104
	*matchp = match;
	return LDAP_SUCCESS;
}

105
/* Index generation function */
106
int octetStringIndexer(
107
108
	slap_mask_t use,
	slap_mask_t flags,
109
110
111
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
112
	BerVarray values,
Howard Chu's avatar
Howard Chu committed
113
114
	BerVarray *keysp,
	void *ctx )
115
116
117
{
	int i;
	size_t slen, mlen;
118
	BerVarray keys;
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
119
120
	HASH_CONTEXT HASHcontext;
	unsigned char HASHdigest[HASH_BYTES];
121
	struct berval digest;
122
123
	digest.bv_val = HASHdigest;
	digest.bv_len = sizeof(HASHdigest);
124

125
	for( i=0; values[i].bv_val != NULL; i++ ) {
126
127
128
		/* just count them */
	}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
129
130
131
	/* we should have at least one value at this point */
	assert( i > 0 );

Howard Chu's avatar
Howard Chu committed
132
	keys = sl_malloc( sizeof( struct berval ) * (i+1), ctx );
133

134
135
	slen = syntax->ssyn_oidlen;
	mlen = mr->smr_oidlen;
136

137
	for( i=0; values[i].bv_val != NULL; i++ ) {
138
		HASH_Init( &HASHcontext );
139
		if( prefix != NULL && prefix->bv_len > 0 ) {
140
			HASH_Update( &HASHcontext,
141
142
				prefix->bv_val, prefix->bv_len );
		}
143
		HASH_Update( &HASHcontext,
144
			syntax->ssyn_oid, slen );
145
		HASH_Update( &HASHcontext,
146
			mr->smr_oid, mlen );
147
		HASH_Update( &HASHcontext,
148
			values[i].bv_val, values[i].bv_len );
149
		HASH_Final( HASHdigest, &HASHcontext );
150

Howard Chu's avatar
Howard Chu committed
151
		ber_dupbv_x( &keys[i], &digest, ctx );
152
153
	}

154
	keys[i].bv_val = NULL;
155
	keys[i].bv_len = 0;
156
157
158
159
160
161
162

	*keysp = keys;

	return LDAP_SUCCESS;
}

/* Index generation function */
163
int octetStringFilter(
164
165
	slap_mask_t use,
	slap_mask_t flags,
166
167
168
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
169
	void * assertedValue,
Howard Chu's avatar
Howard Chu committed
170
171
	BerVarray *keysp,
	void *ctx )
172
173
{
	size_t slen, mlen;
174
	BerVarray keys;
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
175
176
	HASH_CONTEXT HASHcontext;
	unsigned char HASHdigest[HASH_BYTES];
177
	struct berval *value = (struct berval *) assertedValue;
178
	struct berval digest;
179
180
	digest.bv_val = HASHdigest;
	digest.bv_len = sizeof(HASHdigest);
181

182
183
	slen = syntax->ssyn_oidlen;
	mlen = mr->smr_oidlen;
184

Howard Chu's avatar
Howard Chu committed
185
	keys = sl_malloc( sizeof( struct berval ) * 2, ctx );
186

187
	HASH_Init( &HASHcontext );
188
	if( prefix != NULL && prefix->bv_len > 0 ) {
189
		HASH_Update( &HASHcontext,
190
191
			prefix->bv_val, prefix->bv_len );
	}
192
	HASH_Update( &HASHcontext,
193
		syntax->ssyn_oid, slen );
194
	HASH_Update( &HASHcontext,
195
		mr->smr_oid, mlen );
196
	HASH_Update( &HASHcontext,
197
		value->bv_val, value->bv_len );
198
	HASH_Final( HASHdigest, &HASHcontext );
199

Howard Chu's avatar
Howard Chu committed
200
	ber_dupbv_x( keys, &digest, ctx );
201
	keys[1].bv_val = NULL;
202
	keys[1].bv_len = 0;
203
204
205
206
207

	*keysp = keys;

	return LDAP_SUCCESS;
}
208

209
static int
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
210
octetStringSubstringsMatch(
Kurt Zeilenga's avatar
Kurt Zeilenga committed
211
212
	int *matchp,
	slap_mask_t flags,
213
	Syntax *syntax,
Kurt Zeilenga's avatar
Kurt Zeilenga committed
214
215
216
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
217
{
Kurt Zeilenga's avatar
Kurt Zeilenga committed
218
219
220
221
	int match = 0;
	SubstringsAssertion *sub = assertedValue;
	struct berval left = *value;
	int i;
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
222
	ber_len_t inlen = 0;
223

Kurt Zeilenga's avatar
Kurt Zeilenga committed
224
225
226
	/* Add up asserted input length */
	if( sub->sa_initial.bv_val ) {
		inlen += sub->sa_initial.bv_len;
227
	}
Kurt Zeilenga's avatar
Kurt Zeilenga committed
228
229
230
231
232
233
234
	if( sub->sa_any ) {
		for(i=0; sub->sa_any[i].bv_val != NULL; i++) {
			inlen += sub->sa_any[i].bv_len;
		}
	}
	if( sub->sa_final.bv_val ) {
		inlen += sub->sa_final.bv_len;
235
236
	}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
237
238
239
240
	if( sub->sa_initial.bv_val ) {
		if( inlen > left.bv_len ) {
			match = 1;
			goto done;
241
242
		}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
243
244
		match = memcmp( sub->sa_initial.bv_val, left.bv_val,
			sub->sa_initial.bv_len );
245

Kurt Zeilenga's avatar
Kurt Zeilenga committed
246
247
248
		if( match != 0 ) {
			goto done;
		}
249

Kurt Zeilenga's avatar
Kurt Zeilenga committed
250
251
252
253
		left.bv_val += sub->sa_initial.bv_len;
		left.bv_len -= sub->sa_initial.bv_len;
		inlen -= sub->sa_initial.bv_len;
	}
254

Kurt Zeilenga's avatar
Kurt Zeilenga committed
255
256
257
258
259
	if( sub->sa_final.bv_val ) {
		if( inlen > left.bv_len ) {
			match = 1;
			goto done;
		}
260

Kurt Zeilenga's avatar
Kurt Zeilenga committed
261
262
263
		match = memcmp( sub->sa_final.bv_val,
			&left.bv_val[left.bv_len - sub->sa_final.bv_len],
			sub->sa_final.bv_len );
264

Kurt Zeilenga's avatar
Kurt Zeilenga committed
265
266
		if( match != 0 ) {
			goto done;
267
268
		}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
269
270
		left.bv_len -= sub->sa_final.bv_len;
		inlen -= sub->sa_final.bv_len;
271
272
	}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
273
274
275
276
	if( sub->sa_any ) {
		for(i=0; sub->sa_any[i].bv_val; i++) {
			ber_len_t idx;
			char *p;
277

Kurt Zeilenga's avatar
Kurt Zeilenga committed
278
279
280
281
282
283
retry:
			if( inlen > left.bv_len ) {
				/* not enough length */
				match = 1;
				goto done;
			}
284

Kurt Zeilenga's avatar
Kurt Zeilenga committed
285
286
287
			if( sub->sa_any[i].bv_len == 0 ) {
				continue;
			}
288

Kurt Zeilenga's avatar
Kurt Zeilenga committed
289
			p = memchr( left.bv_val, *sub->sa_any[i].bv_val, left.bv_len );
290

Kurt Zeilenga's avatar
Kurt Zeilenga committed
291
292
293
			if( p == NULL ) {
				match = 1;
				goto done;
294
295
			}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
296
			idx = p - left.bv_val;
297

Kurt Zeilenga's avatar
Kurt Zeilenga committed
298
299
300
301
			if( idx >= left.bv_len ) {
				/* this shouldn't happen */
				return LDAP_OTHER;
			}
302

Kurt Zeilenga's avatar
Kurt Zeilenga committed
303
304
			left.bv_val = p;
			left.bv_len -= idx;
305

Kurt Zeilenga's avatar
Kurt Zeilenga committed
306
307
308
309
310
			if( sub->sa_any[i].bv_len > left.bv_len ) {
				/* not enough left */
				match = 1;
				goto done;
			}
311

Kurt Zeilenga's avatar
Kurt Zeilenga committed
312
313
314
			match = memcmp( left.bv_val,
				sub->sa_any[i].bv_val,
				sub->sa_any[i].bv_len );
315

Kurt Zeilenga's avatar
Kurt Zeilenga committed
316
317
318
319
320
			if( match != 0 ) {
				left.bv_val++;
				left.bv_len--;
				goto retry;
			}
321

Kurt Zeilenga's avatar
Kurt Zeilenga committed
322
323
324
			left.bv_val += sub->sa_any[i].bv_len;
			left.bv_len -= sub->sa_any[i].bv_len;
			inlen -= sub->sa_any[i].bv_len;
325
326
327
		}
	}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
328
329
done:
	*matchp = match;
330
331
332
	return LDAP_SUCCESS;
}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
333
334
/* Substrings Index generation function */
static int
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
335
octetStringSubstringsIndexer(
Kurt Zeilenga's avatar
Kurt Zeilenga committed
336
337
338
339
340
341
	slap_mask_t use,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	BerVarray values,
Howard Chu's avatar
Howard Chu committed
342
343
	BerVarray *keysp,
	void *ctx )
Kurt Zeilenga's avatar
Kurt Zeilenga committed
344
345
346
347
348
{
	ber_len_t i, j, nkeys;
	size_t slen, mlen;
	BerVarray keys;

Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
349
350
	HASH_CONTEXT HASHcontext;
	unsigned char HASHdigest[HASH_BYTES];
Kurt Zeilenga's avatar
Kurt Zeilenga committed
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
	struct berval digest;
	digest.bv_val = HASHdigest;
	digest.bv_len = sizeof(HASHdigest);

	nkeys=0;

	for( i=0; values[i].bv_val != NULL; i++ ) {
		/* count number of indices to generate */
		if( values[i].bv_len < SLAP_INDEX_SUBSTR_MINLEN ) {
			continue;
		}

		if( flags & SLAP_INDEX_SUBSTR_INITIAL ) {
			if( values[i].bv_len >= SLAP_INDEX_SUBSTR_MAXLEN ) {
				nkeys += SLAP_INDEX_SUBSTR_MAXLEN -
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
366
					(SLAP_INDEX_SUBSTR_MINLEN - 1);
Kurt Zeilenga's avatar
Kurt Zeilenga committed
367
			} else {
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
368
				nkeys += values[i].bv_len - (SLAP_INDEX_SUBSTR_MINLEN - 1);
Kurt Zeilenga's avatar
Kurt Zeilenga committed
369
370
371
372
373
			}
		}

		if( flags & SLAP_INDEX_SUBSTR_ANY ) {
			if( values[i].bv_len >= SLAP_INDEX_SUBSTR_MAXLEN ) {
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
374
				nkeys += values[i].bv_len - (SLAP_INDEX_SUBSTR_MAXLEN - 1);
Kurt Zeilenga's avatar
Kurt Zeilenga committed
375
376
377
378
379
380
381
382
			}
		}

		if( flags & SLAP_INDEX_SUBSTR_FINAL ) {
			if( values[i].bv_len >= SLAP_INDEX_SUBSTR_MAXLEN ) {
				nkeys += SLAP_INDEX_SUBSTR_MAXLEN -
					( SLAP_INDEX_SUBSTR_MINLEN - 1);
			} else {
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
383
				nkeys += values[i].bv_len - (SLAP_INDEX_SUBSTR_MINLEN - 1);
Kurt Zeilenga's avatar
Kurt Zeilenga committed
384
385
386
387
388
389
390
391
392
393
			}
		}
	}

	if( nkeys == 0 ) {
		/* no keys to generate */
		*keysp = NULL;
		return LDAP_SUCCESS;
	}

Howard Chu's avatar
Howard Chu committed
394
	keys = sl_malloc( sizeof( struct berval ) * (nkeys+1), ctx );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
395
396
397
398
399
400
401
402
403
404
405
406
407
408

	slen = syntax->ssyn_oidlen;
	mlen = mr->smr_oidlen;

	nkeys=0;
	for( i=0; values[i].bv_val != NULL; i++ ) {
		ber_len_t j,max;

		if( values[i].bv_len < SLAP_INDEX_SUBSTR_MINLEN ) continue;

		if( ( flags & SLAP_INDEX_SUBSTR_ANY ) &&
			( values[i].bv_len >= SLAP_INDEX_SUBSTR_MAXLEN ) )
		{
			char pre = SLAP_INDEX_SUBSTR_PREFIX;
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
409
			max = values[i].bv_len - (SLAP_INDEX_SUBSTR_MAXLEN - 1);
Kurt Zeilenga's avatar
Kurt Zeilenga committed
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428

			for( j=0; j<max; j++ ) {
				HASH_Init( &HASHcontext );
				if( prefix != NULL && prefix->bv_len > 0 ) {
					HASH_Update( &HASHcontext,
						prefix->bv_val, prefix->bv_len );
				}

				HASH_Update( &HASHcontext,
					&pre, sizeof( pre ) );
				HASH_Update( &HASHcontext,
					syntax->ssyn_oid, slen );
				HASH_Update( &HASHcontext,
					mr->smr_oid, mlen );
				HASH_Update( &HASHcontext,
					&values[i].bv_val[j],
					SLAP_INDEX_SUBSTR_MAXLEN );
				HASH_Final( HASHdigest, &HASHcontext );

Howard Chu's avatar
Howard Chu committed
429
				ber_dupbv_x( &keys[nkeys++], &digest, ctx );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
			}
		}

		max = SLAP_INDEX_SUBSTR_MAXLEN < values[i].bv_len
			? SLAP_INDEX_SUBSTR_MAXLEN : values[i].bv_len;

		for( j=SLAP_INDEX_SUBSTR_MINLEN; j<=max; j++ ) {
			char pre;

			if( flags & SLAP_INDEX_SUBSTR_INITIAL ) {
				pre = SLAP_INDEX_SUBSTR_INITIAL_PREFIX;
				HASH_Init( &HASHcontext );
				if( prefix != NULL && prefix->bv_len > 0 ) {
					HASH_Update( &HASHcontext,
						prefix->bv_val, prefix->bv_len );
				}
				HASH_Update( &HASHcontext,
					&pre, sizeof( pre ) );
				HASH_Update( &HASHcontext,
					syntax->ssyn_oid, slen );
				HASH_Update( &HASHcontext,
					mr->smr_oid, mlen );
				HASH_Update( &HASHcontext,
					values[i].bv_val, j );
				HASH_Final( HASHdigest, &HASHcontext );

Howard Chu's avatar
Howard Chu committed
456
				ber_dupbv_x( &keys[nkeys++], &digest, ctx );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
			}

			if( flags & SLAP_INDEX_SUBSTR_FINAL ) {
				pre = SLAP_INDEX_SUBSTR_FINAL_PREFIX;
				HASH_Init( &HASHcontext );
				if( prefix != NULL && prefix->bv_len > 0 ) {
					HASH_Update( &HASHcontext,
						prefix->bv_val, prefix->bv_len );
				}
				HASH_Update( &HASHcontext,
					&pre, sizeof( pre ) );
				HASH_Update( &HASHcontext,
					syntax->ssyn_oid, slen );
				HASH_Update( &HASHcontext,
					mr->smr_oid, mlen );
				HASH_Update( &HASHcontext,
					&values[i].bv_val[values[i].bv_len-j], j );
				HASH_Final( HASHdigest, &HASHcontext );

Howard Chu's avatar
Howard Chu committed
476
				ber_dupbv_x( &keys[nkeys++], &digest, ctx );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
			}

		}

	}

	if( nkeys > 0 ) {
		keys[nkeys].bv_val = NULL;
		*keysp = keys;
	} else {
		ch_free( keys );
		*keysp = NULL;
	}

	return LDAP_SUCCESS;
}

static int
octetStringSubstringsFilter (
	slap_mask_t use,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	void * assertedValue,
Howard Chu's avatar
Howard Chu committed
502
503
	BerVarray *keysp,
	void *ctx)
Kurt Zeilenga's avatar
Kurt Zeilenga committed
504
505
506
507
508
509
{
	SubstringsAssertion *sa;
	char pre;
	ber_len_t nkeys = 0;
	size_t slen, mlen, klen;
	BerVarray keys;
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
510
511
	HASH_CONTEXT HASHcontext;
	unsigned char HASHdigest[HASH_BYTES];
Kurt Zeilenga's avatar
Kurt Zeilenga committed
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
	struct berval *value;
	struct berval digest;

	sa = (SubstringsAssertion *) assertedValue;

	if( flags & SLAP_INDEX_SUBSTR_INITIAL && sa->sa_initial.bv_val != NULL
		&& sa->sa_initial.bv_len >= SLAP_INDEX_SUBSTR_MINLEN )
	{
		nkeys++;
	}

	if( flags & SLAP_INDEX_SUBSTR_ANY && sa->sa_any != NULL ) {
		ber_len_t i;
		for( i=0; sa->sa_any[i].bv_val != NULL; i++ ) {
			if( sa->sa_any[i].bv_len >= SLAP_INDEX_SUBSTR_MAXLEN ) {
				/* don't bother accounting for stepping */
				nkeys += sa->sa_any[i].bv_len -
					( SLAP_INDEX_SUBSTR_MAXLEN - 1 );
			}
		}
	}

	if( flags & SLAP_INDEX_SUBSTR_FINAL && sa->sa_final.bv_val != NULL &&
		sa->sa_final.bv_len >= SLAP_INDEX_SUBSTR_MINLEN )
	{
		nkeys++;
	}

	if( nkeys == 0 ) {
		*keysp = NULL;
		return LDAP_SUCCESS;
	}

	digest.bv_val = HASHdigest;
	digest.bv_len = sizeof(HASHdigest);

	slen = syntax->ssyn_oidlen;
	mlen = mr->smr_oidlen;

Howard Chu's avatar
Howard Chu committed
551
	keys = sl_malloc( sizeof( struct berval ) * (nkeys+1), ctx );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
	nkeys = 0;

	if( flags & SLAP_INDEX_SUBSTR_INITIAL && sa->sa_initial.bv_val != NULL &&
		sa->sa_initial.bv_len >= SLAP_INDEX_SUBSTR_MINLEN )
	{
		pre = SLAP_INDEX_SUBSTR_INITIAL_PREFIX;
		value = &sa->sa_initial;

		klen = SLAP_INDEX_SUBSTR_MAXLEN < value->bv_len
			? SLAP_INDEX_SUBSTR_MAXLEN : value->bv_len;

		HASH_Init( &HASHcontext );
		if( prefix != NULL && prefix->bv_len > 0 ) {
			HASH_Update( &HASHcontext,
				prefix->bv_val, prefix->bv_len );
		}
		HASH_Update( &HASHcontext,
			&pre, sizeof( pre ) );
		HASH_Update( &HASHcontext,
			syntax->ssyn_oid, slen );
		HASH_Update( &HASHcontext,
			mr->smr_oid, mlen );
		HASH_Update( &HASHcontext,
			value->bv_val, klen );
		HASH_Final( HASHdigest, &HASHcontext );

Howard Chu's avatar
Howard Chu committed
578
		ber_dupbv_x( &keys[nkeys++], &digest, ctx );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
	}

	if( flags & SLAP_INDEX_SUBSTR_ANY && sa->sa_any != NULL ) {
		ber_len_t i, j;
		pre = SLAP_INDEX_SUBSTR_PREFIX;
		klen = SLAP_INDEX_SUBSTR_MAXLEN;

		for( i=0; sa->sa_any[i].bv_val != NULL; i++ ) {
			if( sa->sa_any[i].bv_len < SLAP_INDEX_SUBSTR_MAXLEN ) {
				continue;
			}

			value = &sa->sa_any[i];

			for(j=0;
				j <= value->bv_len - SLAP_INDEX_SUBSTR_MAXLEN;
				j += SLAP_INDEX_SUBSTR_STEP )
			{
				HASH_Init( &HASHcontext );
				if( prefix != NULL && prefix->bv_len > 0 ) {
					HASH_Update( &HASHcontext,
						prefix->bv_val, prefix->bv_len );
				}
				HASH_Update( &HASHcontext,
					&pre, sizeof( pre ) );
				HASH_Update( &HASHcontext,
					syntax->ssyn_oid, slen );
				HASH_Update( &HASHcontext,
					mr->smr_oid, mlen );
				HASH_Update( &HASHcontext,
					&value->bv_val[j], klen ); 
				HASH_Final( HASHdigest, &HASHcontext );

Howard Chu's avatar
Howard Chu committed
612
				ber_dupbv_x( &keys[nkeys++], &digest, ctx );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
			}
		}
	}

	if( flags & SLAP_INDEX_SUBSTR_FINAL && sa->sa_final.bv_val != NULL &&
		sa->sa_final.bv_len >= SLAP_INDEX_SUBSTR_MINLEN )
	{
		pre = SLAP_INDEX_SUBSTR_FINAL_PREFIX;
		value = &sa->sa_final;

		klen = SLAP_INDEX_SUBSTR_MAXLEN < value->bv_len
			? SLAP_INDEX_SUBSTR_MAXLEN : value->bv_len;

		HASH_Init( &HASHcontext );
		if( prefix != NULL && prefix->bv_len > 0 ) {
			HASH_Update( &HASHcontext,
				prefix->bv_val, prefix->bv_len );
		}
		HASH_Update( &HASHcontext,
			&pre, sizeof( pre ) );
		HASH_Update( &HASHcontext,
			syntax->ssyn_oid, slen );
		HASH_Update( &HASHcontext,
			mr->smr_oid, mlen );
		HASH_Update( &HASHcontext,
			&value->bv_val[value->bv_len-klen], klen );
		HASH_Final( HASHdigest, &HASHcontext );

Howard Chu's avatar
Howard Chu committed
641
		ber_dupbv_x( &keys[nkeys++], &digest, ctx );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
	}

	if( nkeys > 0 ) {
		keys[nkeys].bv_val = NULL;
		*keysp = keys;
	} else {
		ch_free( keys );
		*keysp = NULL;
	}

	return LDAP_SUCCESS;
}

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;
	}

	/*
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
670
671
672
	 * RFC 2252 section 6.3 Bit String
	 *	bitstring = "'" *binary-digit "'B"
	 *	binary-digit = "0" / "1"
Kurt Zeilenga's avatar
Kurt Zeilenga committed
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
	 * example: '0101111101'B
	 */
	
	if( in->bv_val[0] != '\'' ||
		in->bv_val[in->bv_len-2] != '\'' ||
		in->bv_val[in->bv_len-1] != 'B' )
	{
		return LDAP_INVALID_SYNTAX;
	}

	for( i=in->bv_len-3; i>0; i-- ) {
		if( in->bv_val[i] != '0' && in->bv_val[i] != '1' ) {
			return LDAP_INVALID_SYNTAX;
		}
	}

	return LDAP_SUCCESS;
}

static int
nameUIDValidate(
	Syntax *syntax,
	struct berval *in )
{
	int rc;
	struct berval dn;

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

	ber_dupbv( &dn, in );
	if( !dn.bv_val ) return LDAP_OTHER;

	if( dn.bv_val[dn.bv_len-1] == 'B'
		&& dn.bv_val[dn.bv_len-2] == '\'' )
	{
		/* assume presence of optional UID */
		ber_len_t i;

		for(i=dn.bv_len-3; i>1; i--) {
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
712
			if( dn.bv_val[i] != '0' && dn.bv_val[i] != '1' ) {
Kurt Zeilenga's avatar
Kurt Zeilenga committed
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
				break;
			}
		}
		if( dn.bv_val[i] != '\'' || dn.bv_val[i-1] != '#' ) {
			ber_memfree( dn.bv_val );
			return LDAP_INVALID_SYNTAX;
		}

		/* trim the UID to allow use of dnValidate */
		dn.bv_val[i-1] = '\0';
		dn.bv_len = i-1;
	}

	rc = dnValidate( NULL, &dn );

	ber_memfree( dn.bv_val );
	return rc;
}

732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
int
nameUIDPretty(
	Syntax *syntax,
	struct berval *val,
	struct berval *out,
	void *ctx )
{
	assert( val );
	assert( out );


#ifdef NEW_LOGGING
	LDAP_LOG( OPERATION, ARGS, ">>> nameUIDPretty: <%s>\n", val->bv_val, 0, 0 );
#else
	Debug( LDAP_DEBUG_TRACE, ">>> nameUIDPretty: <%s>\n", val->bv_val, 0, 0 );
#endif

	if( val->bv_len == 0 ) {
		ber_dupbv_x( out, val, ctx );

	} else if ( val->bv_len > SLAP_LDAPDN_MAXLEN ) {
		return LDAP_INVALID_SYNTAX;

	} else {
		int rc;
		struct berval dnval = *val;
		struct berval uidval = { 0, NULL };

		if( val->bv_val[val->bv_len-1] == 'B'
			&& val->bv_val[val->bv_len-2] == '\'' )
		{
			uidval.bv_val=strrchr( val->bv_val, '#' );
			if( uidval.bv_val ) {
				dnval.bv_len = uidval.bv_val - dnval.bv_val;
				uidval.bv_len = val->bv_len - dnval.bv_len;

				uidval.bv_len--;
				uidval.bv_val++;
			}
		}

		rc = dnPretty( syntax, &dnval, out, ctx );
		if( rc != LDAP_SUCCESS ) return rc;

		if( uidval.bv_val ) {
			char *tmp = sl_realloc( out->bv_val, out->bv_len + uidval.bv_len + 2, ctx );
			int i, c, got1;
			if( tmp == NULL ) {
				ber_memfree_x( out->bv_val, ctx );
				return LDAP_OTHER;
			}
783
			out->bv_val = tmp;
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
			out->bv_val[out->bv_len++] = '#';

			got1 = uidval.bv_len < sizeof("'0'B"); 
			for(i=0; i<uidval.bv_len; i++) {
				c = uidval.bv_val[i];
				switch(c) {
					case '0':
						if( got1 ) out->bv_val[out->bv_len++] = c;
						break;
					case '1':
						got1 = 1;
					default:
						out->bv_val[out->bv_len++] = c;
				}
			}

			out->bv_val[out->bv_len] = '\0';
		}
	}

#ifdef NEW_LOGGING
	LDAP_LOG( OPERATION, ARGS, "<<< nameUIDPretty: <%s>\n", out->bv_val, 0, 0 );
#else
	Debug( LDAP_DEBUG_TRACE, "<<< nameUIDPretty: <%s>\n", out->bv_val, 0, 0 );
#endif

	return LDAP_SUCCESS;
}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
813
814
815
816
817
818
static int
uniqueMemberNormalize(
	slap_mask_t usage,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *val,
Howard Chu's avatar
Howard Chu committed
819
820
	struct berval *normalized,
	void *ctx )
Kurt Zeilenga's avatar
Kurt Zeilenga committed
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
{
	struct berval out;
	int rc;

	ber_dupbv( &out, val );
	if( out.bv_len != 0 ) {
		struct berval uid = { 0, NULL };

		if( out.bv_val[out.bv_len-1] == 'B'
			&& out.bv_val[out.bv_len-2] == '\'' )
		{
			/* assume presence of optional UID */
			uid.bv_val = strrchr( out.bv_val, '#' );

			if( uid.bv_val == NULL ) {
				free( out.bv_val );
				return LDAP_INVALID_SYNTAX;
			}

			uid.bv_len = out.bv_len - (uid.bv_val - out.bv_val);
			out.bv_len -= uid.bv_len--;

			/* temporarily trim the UID */
			*(uid.bv_val++) = '\0';
		}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
847
		rc = dnNormalize( 0, NULL, NULL, &out, normalized, ctx );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875

		if( rc != LDAP_SUCCESS ) {
			free( out.bv_val );
			return LDAP_INVALID_SYNTAX;
		}

		if( uid.bv_len ) {
			normalized->bv_val = ch_realloc( normalized->bv_val,
				normalized->bv_len + uid.bv_len + sizeof("#") );

			/* insert the separator */
			normalized->bv_val[normalized->bv_len++] = '#';

			/* append the UID */
			AC_MEMCPY( &normalized->bv_val[normalized->bv_len],
				uid.bv_val, uid.bv_len );
			normalized->bv_len += uid.bv_len;

			/* terminate */
			normalized->bv_val[normalized->bv_len] = '\0';
		}

		free( out.bv_val );
	}

	return LDAP_SUCCESS;
}

876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
static int
uniqueMemberMatch(
	int *matchp,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
{
	int match;
	struct berval *asserted = (struct berval *) assertedValue;
	struct berval assertedDN = { 0, NULL };
	struct berval assertedUID = { 0, NULL };
	struct berval valueDN = { 0, NULL };
	struct berval valueUID = { 0, NULL };

	if( asserted->bv_len != 0 ) {
		assertedDN = *asserted;

		if( assertedDN.bv_val[assertedDN.bv_len-1] == 'B'
			&& assertedDN.bv_val[assertedDN.bv_len-2] == '\'' )
		{
			/* assume presence of optional UID */
			assertedUID.bv_val = strrchr( assertedDN.bv_val, '#' );

			if( assertedUID.bv_val == NULL ) {
				return LDAP_INVALID_SYNTAX;
			}

			assertedUID.bv_len = assertedDN.bv_len -
				(assertedUID.bv_val - assertedDN.bv_val);
			assertedDN.bv_len -= assertedUID.bv_len--;

			/* trim the separator */
			assertedUID.bv_val++;
		}
	}

	if( value->bv_len != 0 ) {
		valueDN = *value;

		if( valueDN.bv_val[valueDN.bv_len-1] == 'B'
			&& valueDN.bv_val[valueDN.bv_len-2] == '\'' )
		{
			/* assume presence of optional UID */
			valueUID.bv_val = strrchr( valueDN.bv_val, '#' );

			if( valueUID.bv_val == NULL ) {
				return LDAP_INVALID_SYNTAX;
			}

			valueUID.bv_len = valueDN.bv_len -
				(assertedUID.bv_val - assertedDN.bv_val);
			valueDN.bv_len -= valueUID.bv_len--;

			/* trim the separator */
			valueUID.bv_val++;
		}
	}

	if( valueUID.bv_len && assertedUID.bv_len ) {
		match = memcmp( valueUID.bv_val, assertedUID.bv_val, valueUID.bv_len );
		if( match ) {
			*matchp = match;
			return LDAP_SUCCESS;
		}
	}

	return dnMatch( matchp, flags, syntax, mr, &valueDN, &assertedDN );
}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
947
948
949
950
951
952
/*
 * 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.
 */
953
954
955
956
957
958
959
960
961
962
static int
booleanValidate(
	Syntax *syntax,
	struct berval *in )
{
	/* very unforgiving validation, requires no normalization
	 * before simplistic matching
	 */

	if( in->bv_len == 4 ) {
963
		if( bvmatch( in, &slap_true_bv ) ) {
964
965
966
			return LDAP_SUCCESS;
		}
	} else if( in->bv_len == 5 ) {
967
		if( bvmatch( in, &slap_false_bv ) ) {
968
969
970
971
972
973
974
975
976
977
			return LDAP_SUCCESS;
		}
	}

	return LDAP_INVALID_SYNTAX;
}

static int
booleanMatch(
	int *matchp,
978
	slap_mask_t flags,
979
980
981
982
983
984
985
986
987
988
989
	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;
}

990
991
992
993
994
/*-------------------------------------------------------------------
LDAP/X.500 string syntax / matching rules have a few oddities.  This
comment attempts to detail how slapd(8) treats them.

Summary:
995
  StringSyntax		X.500	LDAP	Matching/Comments
996
997
  DirectoryString	CHOICE	UTF8	i/e + ignore insignificant spaces
  PrintableString	subset	subset	i/e + ignore insignificant spaces
998
  PrintableString	subset	subset	i/e + ignore insignificant spaces
Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
999
  NumericString		subset	subset	ignore all spaces
1000
1001
1002
  IA5String			ASCII	ASCII	i/e + ignore insignificant spaces
  TeletexString		T.61	T.61	i/e + ignore insignificant spaces

Kurt Zeilenga's avatar
cleanup    
Kurt Zeilenga committed
1003
  TelephoneNumber	subset	subset	i + ignore all spaces and "-"
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013

  See draft-ietf-ldapbis-strpro for details (once published).


Directory String -
  In X.500(93), a directory string can be either a PrintableString,
  a bmpString, or a UniversalString (e.g., UCS (a subset of Unicode)).
  In later versions, more CHOICEs were added.  In all cases the string
  must be non-empty.

1014
  In LDAPv3, a directory string is a UTF-8 encoded UCS string.
1015
  A directory string cannot be zero length.
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061

  For matching, there are both case ignore and exact rules.  Both
  also require that "insignificant" spaces be ignored.
	spaces before the first non-space are ignored;
	spaces after the last non-space are ignored;
	spaces after a space are ignored.
  Note: by these rules (and as clarified in X.520), a string of only
  spaces is to be treated as if held one space, not empty (which
  would be a syntax error).

NumericString
  In ASN.1, numeric string is just a string of digits and spaces
  and could be empty.  However, in X.500, all attribute values of
  numeric string carry a non-empty constraint.  For example:

	internationalISDNNumber ATTRIBUTE ::= {
		WITH SYNTAX InternationalISDNNumber
		EQUALITY MATCHING RULE numericStringMatch
		SUBSTRINGS MATCHING RULE numericStringSubstringsMatch
		ID id-at-internationalISDNNumber }
	InternationalISDNNumber ::=
	    NumericString (SIZE(1..ub-international-isdn-number))

  Unforunately, some assertion values are don't carry the same
  constraint (but its unclear how such an assertion could ever
  be true). In LDAP, there is one syntax (numericString) not two
  (numericString with constraint, numericString without constraint).
  This should be treated as numericString with non-empty constraint.
  Note that while someone may have no ISDN number, there are no ISDN
  numbers which are zero length.

  In matching, spaces are ignored.

PrintableString
  In ASN.1, Printable string is just a string of printable characters
  and can be empty.  In X.500, semantics much like NumericString (see
  serialNumber for a like example) excepting uses insignificant space
  handling instead of ignore all spaces.  

IA5String
  Basically same as PrintableString.  There are no examples in X.500,
  but same logic applies.  So we require them to be non-empty as
  well.

-------------------------------------------------------------------*/

1062
1063
1064
1065
1066
1067
1068
1069
1070
static int
UTF8StringValidate(
	Syntax *syntax,
	struct berval *in )
{
	ber_len_t count;
	int len;
	unsigned char *u = in->bv_val;

1071
1072
1073
1074
	if( in->bv_len == 0 && syntax == slap_schema.si_syn_directoryString ) {
		/* directory strings cannot be empty */
		return LDAP_INVALID_SYNTAX;
	}
1075

1076
	for( count = in->bv_len; count > 0; count-=len, u+=len ) {
1077
		/* get the length indicated by the first byte */
1078
		len = LDAP_UTF8_CHARLEN2( u, len );
1079

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1080
1081
1082
		/* very basic checks */
		switch( len ) {
			case 6:
1083
				if( (u[5] & 0xC0) != 0x80 ) {
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1084
1085
1086
					return LDAP_INVALID_SYNTAX;
				}
			case 5:
1087
				if( (u[4] & 0xC0) != 0x80 ) {
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1088
1089
1090
					return LDAP_INVALID_SYNTAX;
				}
			case 4:
1091
				if( (u[3] & 0xC0) != 0x80 ) {
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1092
1093
1094
					return LDAP_INVALID_SYNTAX;
				}
			case 3:
1095
				if( (u[2] & 0xC0 )!= 0x80 ) {
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1096
1097
1098
					return LDAP_INVALID_SYNTAX;
				}
			case 2:
1099
				if( (u[1] & 0xC0) != 0x80 ) {
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1100
1101
1102
					return LDAP_INVALID_SYNTAX;
				}
			case 1:
1103
				/* CHARLEN already validated it */
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1104
1105
1106
1107
				break;
			default:
				return LDAP_INVALID_SYNTAX;
		}
1108
1109
1110

		/* make sure len corresponds with the offset
			to the next character */
1111
		if( LDAP_UTF8_OFFSET( u ) != len ) return LDAP_INVALID_SYNTAX;
1112
1113
	}

1114
1115
1116
	if( count != 0 ) {
		return LDAP_INVALID_SYNTAX;
	}
1117

1118
	return LDAP_SUCCESS;
1119
1120
}

1121
1122
1123
1124
1125
1126
static int
UTF8StringNormalize(
	slap_mask_t use,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *val,
Howard Chu's avatar
Howard Chu committed
1127
1128
	struct berval *normalized,
	void *ctx )
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
{
	struct berval tmp, nvalue;
	int flags;
	int i, wasspace;

	if( val->bv_val == NULL ) {
		/* assume we're dealing with a syntax (e.g., UTF8String)
		 * which allows empty strings
		 */
		normalized->bv_len = 0;
		normalized->bv_val = NULL;
		return LDAP_SUCCESS;
	}

1143
	flags = SLAP_MR_ASSOCIATED( mr, slap_schema.si_mr_caseExactMatch )
1144
		? LDAP_UTF8_NOCASEFOLD : LDAP_UTF8_CASEFOLD;
1145
	flags |= ( ( use & SLAP_MR_EQUALITY_APPROX ) == SLAP_MR_EQUALITY_APPROX )
1146
1147
		? LDAP_UTF8_APPROX : 0;

Howard Chu's avatar
Howard Chu committed
1148
	val = UTF8bvnormalize( val, &tmp, flags, ctx );
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
	if( val == NULL ) {
		return LDAP_OTHER;
	}
	
	/* collapse spaces (in place) */
	nvalue.bv_len = 0;
	nvalue.bv_val = tmp.bv_val;

	wasspace=1; /* trim leading spaces */
	for( i=0; i<tmp.bv_len; i++) {
		if ( ASCII_SPACE( tmp.bv_val[i] )) {
			if( wasspace++ == 0 ) {
				/* trim repeated spaces */
				nvalue.bv_val[nvalue.bv_len++] = tmp.bv_val[i];
			}
		} else {
			wasspace = 0;
			nvalue.bv_val[nvalue.bv_len++] = tmp.bv_val[i];
		}
	}

	if( nvalue.bv_len ) {
		if( wasspace ) {
			/* last character was a space, trim it */
			--nvalue.bv_len;
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1174
1175
		}
		nvalue.bv_val[nvalue.bv_len] = '\0';
1176

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1177
1178
1179
1180
1181
1182
	} else {
		/* string of all spaces is treated as one space */
		nvalue.bv_val[0] = ' ';
		nvalue.bv_val[1] = '\0';
		nvalue.bv_len = 1;
	}
1183

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1184
	*normalized = nvalue;
1185
1186
1187
	return LDAP_SUCCESS;
}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1188
#if defined(SLAPD_APPROX_INITIALS)
1189
1190
#	define SLAPD_APPROX_DELIMITER "._ "
#	define SLAPD_APPROX_WORDLEN 2
1191
#else
1192
1193
#	define SLAPD_APPROX_DELIMITER " "
#	define SLAPD_APPROX_WORDLEN 1
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1194
#endif
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204

static int
approxMatch(
	int *matchp,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *value,
	void *assertedValue )
{
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1205
1206
1207
	struct berval *nval, *assertv;
	char *val, **values, **words, *c;
	int i, count, len, nextchunk=0, nextavail=0;
1208

1209
	/* Yes, this is necessary */
1210
	nval = UTF8bvnormalize( value, NULL, LDAP_UTF8_APPROX, NULL );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1211
	if( nval == NULL ) {
1212
1213
1214
1215
1216
		*matchp = 1;
		return LDAP_SUCCESS;
	}

	/* Yes, this is necessary */
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1217
	assertv = UTF8bvnormalize( ((struct berval *)assertedValue),
1218
		NULL, LDAP_UTF8_APPROX, NULL );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1219
1220
1221
	if( assertv == NULL ) {
		ber_bvfree( nval );
		*matchp = 1;
1222
1223
1224
		return LDAP_SUCCESS;
	}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1225
1226
1227
1228
1229
1230
	/* Isolate how many words there are */
	for ( c = nval->bv_val, count = 1; *c; c++ ) {
		c = strpbrk( c, SLAPD_APPROX_DELIMITER );
		if ( c == NULL ) break;
		*c = '\0';
		count++;
1231
1232
	}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1233
1234
1235
1236
1237
1238
	/* Get a phonetic copy of each word */
	words = (char **)ch_malloc( count * sizeof(char *) );
	values = (char **)ch_malloc( count * sizeof(char *) );
	for ( c = nval->bv_val, i = 0;  i < count; i++, c += strlen(c) + 1 ) {
		words[i] = c;
		values[i] = phonetic(c);
1239
	}
1240

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1241
1242
1243
1244
1245
1246
1247
	/* Work through the asserted value's words, to see if at least some
	   of the words are there, in the same order. */
	len = 0;
	while ( (ber_len_t) nextchunk < assertv->bv_len ) {
		len = strcspn( assertv->bv_val + nextchunk, SLAPD_APPROX_DELIMITER);
		if( len == 0 ) {
			nextchunk++;
1248
			continue;
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1249
		}
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1250
1251
1252
1253
1254
1255
1256
1257
#if defined(SLAPD_APPROX_INITIALS)
		else if( len == 1 ) {
			/* Single letter words need to at least match one word's initial */
			for( i=nextavail; i<count; i++ )
				if( !strncasecmp( assertv->bv_val + nextchunk, words[i], 1 )) {
					nextavail=i+1;
					break;
				}
1258
		}
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1259
1260
1261
1262
1263
#endif
		else {
			/* Isolate the next word in the asserted value and phonetic it */
			assertv->bv_val[nextchunk+len] = '\0';
			val = phonetic( assertv->bv_val + nextchunk );
1264

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1265
1266
1267
1268
1269
1270
			/* 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;
				}
1271
			}
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1272
			ch_free( val );
1273
		}
1274

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1275
1276
1277
1278
		/* This chunk in the asserted value was NOT within the *value. */
		if( i >= count ) {
			nextavail=-1;
			break;
1279
1280
		}

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1281
1282
		/* Go on to the next word in the asserted value */
		nextchunk += len+1;
1283
	}
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1284

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1285
1286
1287
1288
1289
1290
1291
	/* If some of the words were seen, call it a match */
	if( nextavail > 0 ) {
		*matchp = 0;
	}
	else {
		*matchp = 1;
	}
1292

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1293
1294
1295
1296
1297
1298
1299
1300
	/* Cleanup allocs */
	ber_bvfree( assertv );
	for( i=0; i<count; i++ ) {
		ch_free( values[i] );
	}
	ch_free( values );
	ch_free( words );
	ber_bvfree( nval );
1301

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1302
1303
	return LDAP_SUCCESS;
}
1304

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1305
1306
1307
1308
1309
1310
1311
1312
static int 
approxIndexer(
	slap_mask_t use,
	slap_mask_t flags,
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
	BerVarray values,
1313
1314
	BerVarray *keysp,
	void *ctx )
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1315
1316
1317
1318
1319
{
	char *c;
	int i,j, len, wordcount, keycount=0;
	struct berval *newkeys;
	BerVarray keys=NULL;
1320

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1321
1322
1323
	for( j=0; values[j].bv_val != NULL; j++ ) {
		struct berval val = { 0, NULL };
		/* Yes, this is necessary */
1324
		UTF8bvnormalize( &values[j], &val, LDAP_UTF8_APPROX, NULL );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1325
		assert( val.bv_val != NULL );
1326

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1327
1328
1329
1330
1331
1332
1333
1334
		/* Isolate how many words there are. There will be a key for each */
		for( wordcount = 0, c = val.bv_val; *c; c++) {
			len = strcspn(c, SLAPD_APPROX_DELIMITER);
			if( len >= SLAPD_APPROX_WORDLEN ) wordcount++;
			c+= len;
			if (*c == '\0') break;
			*c = '\0';
		}
1335

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1336
1337
1338
1339
1340
1341
		/* Allocate/increase storage to account for new keys */
		newkeys = (struct berval *)ch_malloc( (keycount + wordcount + 1) 
			* sizeof(struct berval) );
		AC_MEMCPY( newkeys, keys, keycount * sizeof(struct berval) );
		if( keys ) ch_free( keys );
		keys = newkeys;
1342

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1343
1344
1345
1346
1347
1348
1349
		/* Get a phonetic copy of each word */
		for( c = val.bv_val, i = 0; i < wordcount; c += len + 1 ) {
			len = strlen( c );
			if( len < SLAPD_APPROX_WORDLEN ) continue;
			ber_str2bv( phonetic( c ), 0, 0, &keys[keycount] );
			keycount++;
			i++;
1350
		}
1351

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1352
		ber_memfree( val.bv_val );
1353
	}
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1354
1355
	keys[keycount].bv_val = NULL;
	*keysp = keys;
1356

1357
1358
	return LDAP_SUCCESS;
}
1359

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1360
1361
static int 
approxFilter(
1362
	slap_mask_t use,
1363
	slap_mask_t flags,
1364
1365
1366
	Syntax *syntax,
	MatchingRule *mr,
	struct berval *prefix,
1367
	void * assertedValue,
1368
1369
	BerVarray *keysp,
	void *ctx )
1370
{
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1371
1372
1373
	char *c;
	int i, count, len;
	struct berval *val;
1374
	BerVarray keys;
1375

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1376
1377
	/* Yes, this is necessary */
	val = UTF8bvnormalize( ((struct berval *)assertedValue),
1378
		NULL, LDAP_UTF8_APPROX, NULL );
Kurt Zeilenga's avatar
Kurt Zeilenga committed
1379
1380
1381
1382
1383
1384
	if( val == NULL || val->bv_val == NULL ) {
		keys = (struct berval *)ch_malloc( sizeof(struct berval) );
		keys[0].bv_val = NULL;
		*keysp = keys;
		ber_bvfree( val );
		return LDAP_SUCCESS;
1385
	}
1386

Kurt Zeilenga's avatar
Kurt Zeilenga committed
1387
1388
1389
1390
1391
1392
1393
	/* Isolate how many words there are. There will be a key for each */
	for( count = 0,c = val->bv_val; *c; c++) {
		len = strcspn(c, SLAPD_APPROX_DELIMITER);
		if( len >= SLAPD_APPROX_WORDLEN ) count++;
		c+= len;
		if (*c == '\0') break;
		*c = '\0';
1394
1395
	}

Kurt Zeilenga's avatar