Home | History | Annotate | Line # | Download | only in string 
      1 /*	$NetBSD: wcscspn_bloom.h,v 1.4 2011/11/25 17:48:22 joerg Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 2011 Joerg Sonnenberger,
      5  * All rights reserved.
      6  *
      7  * Redistribution and use in source and binary forms, with or without
      8  * modification, are permitted provided that the following conditions
      9  * are met:
     10  * 1. Redistributions of source code must retain the above copyright
     11  *    notice, this list of conditions and the following disclaimer.
     12  * 2. Redistributions in binary form must reproduce the above copyright
     13  *    notice, this list of conditions and the following disclaimer in the
     14  *    documentation and/or other materials provided with the distribution.
     15  *
     16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
     17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     19  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
     20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     26  * SUCH DAMAGE.
     27  */
     28 
     29 /*
     30  * Bloom filter for fast test if a given character is part of the reject set.
     31  * The test may have false positives, but doesn't have false negatives.
     32  * The first hash function is designed to be very fast to evaluate.
     33  * It is collision free if the input is part of the same European language
     34  * and shouldn't be too bad even other input.  The second hash function
     35  * tries to provide a much better mixing, but involves the slower
     36  * multiplication.
     37  */
     38 
     39 #include <limits.h>
     40 
     41 #define	BLOOM_SIZE		64
     42 #define	BLOOM_ARRAY_SIZE	(BLOOM_SIZE / sizeof(size_t))
     43 #define	BLOOM_BITS		(BLOOM_SIZE * CHAR_BIT)
     44 #define	BLOOM_DIV		(sizeof(size_t) * CHAR_BIT)
     45 
     46 static inline size_t
     47 wcscspn_bloom1(size_t x)
     48 {
     49 	return x % BLOOM_BITS;
     50 }
     51 
     52 static inline size_t
     53 wcscspn_bloom2(size_t x)
     54 {
     55 	return (size_t)((uint32_t)(x * 2654435761U) /
     56 	    (0x100000000ULL / BLOOM_BITS));
     57 }
     58 
     59 static inline void
     60 wcsspn_bloom_init(size_t *bloom, const wchar_t *charset)
     61 {
     62 	size_t val;
     63 
     64 	memset(bloom, 0, BLOOM_SIZE);
     65 	do {
     66 		val = wcscspn_bloom1((size_t)*charset);
     67 		bloom[val / BLOOM_DIV] |= (size_t)1 << (val % BLOOM_DIV);
     68 		val = wcscspn_bloom2((size_t)*charset);
     69 		bloom[val / BLOOM_DIV] |= (size_t)1 << (val % BLOOM_DIV);
     70 	}
     71 	while (*++charset);
     72 }
     73 
     74 static inline int
     75 wcsspn_in_bloom(const size_t *bloom, wchar_t ch)
     76 {
     77 	size_t val;
     78 
     79 	val = wcscspn_bloom1((size_t)ch);
     80 	if (bloom[val / BLOOM_DIV] & ((size_t)1 << (val % BLOOM_DIV)))
     81 		return 1;
     82 	val = wcscspn_bloom2((size_t)ch);
     83 	if (bloom[val / BLOOM_DIV] & ((size_t)1 << (val % BLOOM_DIV)))
     84 		return 1;
     85 	return 0;
     86 }
     87