Mark Dickinson added the comment:
[MAL]
> I don't understand why we are only trying to fix the string problem
> and completely ignore other key types.
[Armin]
> estimating the risks of giving up on a valid query for a truly random
> hash, at an overestimated one billion queries per second ...
That's fine in principle, but if this gets extended to integers, note that our
current integer hash is about as far from 'truly random' as you can get:
Python 3.4.0a0 (default:f02555353544, Nov 4 2012, 11:50:12)
[GCC 4.2.1 (Apple Inc. build 5664)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> [hash(i) for i in range(20)]
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
Moreover, it's going to be *very* hard to change the int hash while preserving
the `x == y implies hash(x) == hash(y)` invariant across all the numeric types
(int, float, complex, Decimal, Fraction, 3rd-party types that need to remain
compatible).
----------
nosy: +mark.dickinson
_______________________________________
Python tracker <[email protected]>
<http://bugs.python.org/issue14621>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
