Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:95324
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.161.193 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <37f541c321ec5cae15d8ca0a1f9c0716@k-piste.dy.fi>
References: <3d8be1faf8d4d3e5add3768d7a8eb16c@k-piste.dy.fi>
 <20160814211942.16358-1-lauri.kentta@gmail.com> <c8095e3b-14b0-3860-4cd9-859c1f5c10a8@thefsb.org>
 <37f541c321ec5cae15d8ca0a1f9c0716@k-piste.dy.fi>
Date: Fri, 19 Aug 2016 20:41:26 +0200
Message-ID: <CAF+90c9MYNZ4H1UepqfznBrDAhv768x_xY6p637oRzECrfe0Yg@mail.gmail.com>
To: =?UTF-8?Q?Lauri_Kentt=C3=A4?= <lauri.kentta@gmail.com>
Cc: PHP internals <internals@lists.php.net>, Tom Worster <fsb@thefsb.org>
Content-Type: multipart/alternative; boundary=94eb2c0779cecc231e053a710be6
Subject: Re: [PHP-DEV] Re: [RFC: PATCH v1] Implement mt_srand_array
From: nikita.ppv@gmail.com (Nikita Popov)

--94eb2c0779cecc231e053a710be6
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

On Fri, Aug 19, 2016 at 8:20 PM, Lauri Kentt=C3=A4 <lauri.kentta@gmail.com>
wrote:

> On 2016-08-15 18:53, Tom Worster wrote:
>
>> Hi Lauri,
>>
>> Do you have a PR against php-src on github? It's easier to read and
>> comment.
>>
>> Tom
>>
>
>
> Here's a PR with a revised patch (also accepts other than pure integers):
>
> https://github.com/php/php-src/pull/2089
>
>
> And here's an alternative with no sanity checks at all, which makes the
> patch even simpler (and more PHP-like, since 'foo' is silently converted =
to
> zero; this truly the most marvelous feature of PHP).
>
> https://github.com/Metabolix/php-src/tree/mt_srand_array-v3
>

Personally I'm not a fan of this. We should not perpetuate the use of
mt_rand() as either a seedable random number generator or an implementation
of a specific algorithm. We have broken the former in 7.1 (seed sequence
change) and were close to breaking the latter (replacing mt19937 with
xorshift128+).

Instead, let's first introduce an object-based PRNG interface, which a)
does not rely on volatile global state and b) is explicit about the choice
of algorithm and seeding procedure.

Thanks,
Nikita

--94eb2c0779cecc231e053a710be6--