Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:96936
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.161.196 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <9c4e33c3-cd8c-dc29-89d2-c91abfb702e0@gmx.de>
References: <2b46dcb5-03e2-59dd-b04f-d1877fd603bc@gmx.de> <CANUQDCjf1cDYzfW3kJ08HLTegccGjo-wZ_6Dwdkco-7i_vb9bw@mail.gmail.com>
 <CAHE1EdsaBdjGuGsiVTPpM=k+A4MUpZBnF68fUCcvUqzg96yXRA@mail.gmail.com> <9c4e33c3-cd8c-dc29-89d2-c91abfb702e0@gmx.de>
Date: Wed, 16 Nov 2016 22:02:45 +0100
Message-ID: <CAF+90c-S-+T=m=1UYfEVeGAXSQJmKyMWVGkfzH-R9icFwhaj5A@mail.gmail.com>
To: "Christoph M. Becker" <cmbecker69@gmx.de>
Cc: Craig Duncan <php@duncanc.co.uk>, Niklas Keller <me@kelunik.com>, 
	Internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary=001a114da1500b73da0541716587
Subject: Re: [PHP-DEV] How to detect classes with a count_elements handler in userland?
From: nikita.ppv@gmail.com (Nikita Popov)

--001a114da1500b73da0541716587
Content-Type: text/plain; charset=UTF-8

On Mon, Nov 14, 2016 at 2:22 PM, Christoph M. Becker <cmbecker69@gmx.de>
wrote:

> On 13.11.2016 at 22:10, Craig Duncan wrote:
>
> >> How about just making those classes implement the interface instead?
> >
> > Christoph pointed out that there may be classes in extensions that use
> > count_elements.
>
> Furthermore, letting those classes implement Countable could break BC.
>

How would this break BC?

SXE even already has a count() method, and count_elements respects
overrides of that method. Clearly SXE is supposed to implement Countable,
it was simply forgotten.

> Also it would make userland code simpler:
> > is_countable($thing)
> > vs
> > is_array($thing) || $thing implements \Countable
>
> Good catch!
>
> Note that Craig has added is_countable() to PR #2185 (which implements
> the RFC).  In my opinion, that is okay, but if anybody has objections,
> we probably need another RFC.
>

This should be done as a separate change IMHO.

Nikita

--001a114da1500b73da0541716587--