diff Objects/setobject.c @ 32320:87302ac07255 legacy-trunk

Futher improvements to frozenset hashing (based on Yitz Gale's battery of tests which nicely highly highlight weaknesses). * Initial value is now a large prime. * Pre-multiply by the set length to add one more basis of differentiation. * Work a bit harder inside the loop to scatter bits from sources that may have closely spaced hash values. All of this is necessary to make up for keep the hash function commutative. Fortunately, the hash value is cached so the call to frozenset_hash() will only occur once per set.
author Raymond Hettinger <python@rcn.com>
date Thu, 10 Jun 2004 22:41:48 +0000
parents b9de6b665273
children a627d9493149
line wrap: on
line diff
--- a/Objects/setobject.c	Thu Jun 10 21:38:41 2004 +0000
+++ b/Objects/setobject.c	Thu Jun 10 22:41:48 2004 +0000
@@ -663,20 +663,22 @@
 	PySetObject *so = (PySetObject *)self;
 	PyObject *key, *value;
 	int pos = 0;
-	long hash = 1905176217L;
+	long hash = 1927868237L;
 
 	if (so->hash != -1)
 		return so->hash;
 
+	hash *= (PyDict_Size(so->data) + 1);
 	while (PyDict_Next(so->data, &pos, &key, &value)) {
-		/* Multiplying by a large prime increases the bit dispersion for
-		   closely spaced hash values.  The is important because some
-		   use cases have many combinations of a small number of 
-		   elements with nearby hashes so that many distinct combinations
-		   collapse to only a handful of distinct hash values. */
-		hash ^= PyObject_Hash(key) * 3644798167u;
+		/* Work to increase the bit dispersion for closely spaced hash
+		   values.  The is important because some use cases have many 
+		   combinations of a small number of elements with nearby 
+		   hashes so that many distinct combinations collapse to only 
+		   a handful of distinct hash values. */
+		long h = PyObject_Hash(key);
+		hash ^= (h ^ (h << 16) ^ 89869747L)  * 3644798167u;
 	}
-	hash *= 69069L;
+	hash = hash * 69069L + 907133923L;
 	if (hash == -1)
 		hash = 590923713L;
 	so->hash = hash;