wn_cumhashu3

Usage

Overview

A <hash function> returns a pseudo-random integer which depends on a <key> argument. The hash functions provided here are very fast and produce very random looking results.

wn_strhash hashes a string.

wn_ptrhash hashes a pointer.

wn_inthash hashes an int.

wn_memhash hashes a memory block of size len. One can use it to hash structures.

The above functions can be assembled to produce hash functions for more complicated data structures, possibly using wn_cumhasho and wn_cumhashu as glue.

wn_cumhasho is designed to produce a joint hash value from several base hash values. The base hash values are passed in hash; the result is accumulated in <accum>. The resulting hash value DEPENDS very heavily on the ORDER of accumulation.

wn_cumhashu is designed to produce a joint hash value from several base hash values. The base hash values are passed in hash; the result is accumulated in <accum>. The resulting hash value does NOT DEPEND on the ORDER of accumulation.

If you are hashing together objects of different types, use for glue the C bit exclusive-or operator <^> instead of the above functions.

Examples

The following function produces a hash value for a list of strings. The result depends very heavily on the order of the list.

int string_list_hash(list)
wn_sll list;
{
wn_sll el;
int hash;
char *string;
hash = 0;
for(el = list;el != NULL;el=el->next)
{
string = el->contents;
wn_cumhasho(&hash,wn_strhash(string));
}
return(hash);
}

The following function produces a hash value for a set of strings. The result does not depend on the order in which the hash calls happen to occur.

int string_set_hash(wn_sll list)
wn_sll list;
{
wn_sll el;
int hash;
char *string;
hash = 0;
for(el=list;el!=NULL;el=el->next)
{
string = el->contents;
wn_cumhashu(&hash,wn_strhash(string));
}
return(hash);
}