Author ygale
Recipients
Date 2004-05-16.12:25:40
SpamBayes Score
Marked as misclassified
Message-id
In-reply-to
Content
Logged In: YES 
user_id=1033539

Hmm, you are correct. This is appears to be
an off-by-one problem: the original seed always gets
multiplied by the constant (which is silly), and the last
item in the tuple does not get multiplied (which causes
the bug).

The correct solution is to change:

value = my_mul(const, value) ^ hash(item)

in Steve's pseudo-code to:

value = my_mul(const, value ^ hash(item))

Of course, you still get a lot more robustness
for almost no cost if you vary "const" across
the tuple via a table.
History
Date User Action Args
2007-08-23 14:21:13adminlinkissue942952 messages
2007-08-23 14:21:13admincreate