Path: csiph.com!usenet.pasdenom.info!gegeweb.org!usenet-fr.net!nerim.net!novso.com!newsfeed.xs4all.nl!newsfeed4a.news.xs4all.nl!xs4all!post.news.xs4all.nl!not-for-mail
MIME-Version: 1.0
In-Reply-To: <1416591555.652129.193873581.54A97BD7@webmail.messagingengine.com>
References: <931d12c0-12b3-4278-a9b9-1e2c32f138b3@googlegroups.com> <87bno16nsw.fsf@elektro.pacujo.net> <1416591555.652129.193873581.54A97BD7@webmail.messagingengine.com>
From: Ian Kelly <ian.g.kelly@gmail.com>
Date: Fri, 21 Nov 2014 11:37:29 -0700
Subject: Re: Infinitely nested containers
To: Python <python-list@python.org>
Content-Type: text/plain; charset=UTF-8
Precedence: list
Newsgroups: comp.lang.python
Message-ID: <mailman.16119.1416595105.18130.python-list@python.org>
Lines: 32
NNTP-Posting-Host: 2001:888:2000:d::a6
Xref: csiph.com comp.lang.python:81236

On Fri, Nov 21, 2014 at 10:39 AM,  <random832@fastmail.us> wrote:
> On Fri, Nov 21, 2014, at 02:00, Marko Rauhamaa wrote:
>> Gill Shen <gillarete@gmail.com>:
>>
>> > How is this [nesting] behavior implemented under the hood?
>>
>> Pointers.
>>
>> > And why is this allowed at all?
>>
>> There's no reason not to.
>
> There's no reason not to allow it with tuples, but you can't do it.
> Mainly because doing it in a single literal would require special
> syntax, whereas you can simply append to a list a reference to itself.
>
> I think I tried on at least one python version and printing the tuple
> crashed with a recursion depth error, since it had no special protection
> for this case the way list printing does.

Here's a nice crash. I thought this might similarly produce a
recursion depth error, but no, it's a seg fault!

$ cat test.py
import itertools

l = []
it = itertools.chain.from_iterable(l)
l.append(it)
next(it)
$ python3 test.py
Segmentation fault (core dumped)