Author rhettinger
Recipients
Date 2004-04-28.16:36:13
SpamBayes Score
Marked as misclassified
Message-id
In-reply-to
Content
Logged In: YES 
user_id=80475

The OP was not referring to "some collisions"; his app
collapsed all entries to a single hash value.  Changing XOR
to + would partially eliminate the self cancelling property
of this hash function. 

Also, I am concerned about the tuple hash using the same
multiplier as the hash for other objects. In sets.py, a
naive combination of the component hash values caused many
distinct sets to collapse to a handful of possibilities --
while tuples do not have an identical issue, it does
highlight the risks involved.

History
Date User Action Args
2007-08-23 14:21:12adminlinkissue942952 messages
2007-08-23 14:21:12admincreate