Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:98573
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain kelunik.com from 81.169.146.221 cause and error)
MIME-Version: 1.0
In-Reply-To: <ba9f19e2-813e-d704-75d9-796f9b634ffa@php.net>
References: <ba9f19e2-813e-d704-75d9-796f9b634ffa@php.net>
Date: Fri, 17 Mar 2017 07:40:45 +0100
Message-ID: <CANUQDCjgRw-p9SVVLevhgi4g7FvQOWi0a8G+R5dyoz1sC5nSbA@mail.gmail.com>
To: PHP Internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary=001a113a6e3412adc4054ae775bf
Subject: Re: [PHP-DEV] Let range() return a generator?
From: me@kelunik.com (Niklas Keller)

--001a113a6e3412adc4054ae775bf
Content-Type: text/plain; charset=UTF-8

>
> For me,
>
>     foreach (range($low, $high) as $i)
>
> is easier to understand than
>
>     for ($i = 0; $i <= $high; $i++)
>
> By easier to understand I do not only mean that the beginning of the loop
> marked by the foreach/range construct is easier to read. While reading the
> body of the loop I can be sure that $i is not unexpectedly modified, for
> instance.
>
> Be that as it may, what I would like to propose should be useful for use
> cases other than the one (see above) that triggered me thinking about this.
>
> range() returns an array that contains all elements requested. Depending
> on the amount of elements, this might not be optimal.
>
> One of the most common usages of generators that I have seen is an
> alternative, userland implementation of the range() function:
>
>     function xrange($start, $end, $step = 1) {
>         for ($i = $start; $i <= $end; $i += $step) {
>             yield $i;
>         }
>     }
>
> This makes me wonder: would there be a benefit in changing the
> implementation of range() to return a generator instead of an array? And if
> so, would this constitute a break in backward compatibility?


Yes, that's a BC break. Instead of changing it, there could be a new
function. But as it can be simply built in userland, I don't see much
reason to have it in core.

Regards, Niklas

--001a113a6e3412adc4054ae775bf--