Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.python > #30501

Re: [RELEASED] Python 3.3.0

References <5066E72E.2010100@python.org> <DUB117-W380D12F6F3E5062720F59491810@phx.gbl> <k46v8h$kcr$1@ger.gmane.org> <k46vue$p6i$1@ger.gmane.org>
Date 2012-09-30 00:25 +1000
Subject Re: [RELEASED] Python 3.3.0
From Chris Angelico <rosuav@gmail.com>
Newsgroups comp.lang.python
Message-ID <mailman.1634.1348928706.27098.python-list@python.org> (permalink)

Show all headers | View raw


On Sun, Sep 30, 2012 at 12:17 AM, Stefan Behnel <stefan_ml@behnel.de> wrote:
> Christian Heimes, 29.09.2012 16:06:
>> From now on you can't rely
>> on the order of an unordered type like dict or set.
>
> Tautologies tend to be true even without a temporal qualification.

Technically people shouldn't ever have relied on the order, but until
hash randomization came in, the order in CPython was actually
predictable - for a given set of dictionary operations, the internal
structure was determinate, and the consequent iteration order would be
consistent. Now that that's no longer the case, "unordered" really
means "unordered", and the order can change from one run of a program
to another.

So it may be a tautology in theory, but not in practice..... until
now. Truth has become true.

ChrisA

Back to comp.lang.python | Previous | Next | Find similar | Unroll thread


Thread

Re: [RELEASED] Python 3.3.0 Chris Angelico <rosuav@gmail.com> - 2012-09-30 00:25 +1000

csiph-web