Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:115283
Content-Type: text/plain;
	charset=utf-8
Content-Transfer-Encoding: quoted-printable
Mime-Version: 1.0 (Mac OS X Mail 11.5 \(3445.9.7\))
Date: Sat, 3 Jul 2021 16:04:27 +0200
References: <0BBAA2F7-02D7-4AEE-B48A-97436A0D8E5D@gmail.com>
 <CAF+90c-X+cXZaoc8oNrsH4ftZRFhf+xSqyc+eMttMT0Roq3u4A@mail.gmail.com>
To: PHP internals <internals@lists.php.net>
In-Reply-To: <CAF+90c-X+cXZaoc8oNrsH4ftZRFhf+xSqyc+eMttMT0Roq3u4A@mail.gmail.com>
Message-ID: <19102874-12E9-448F-AC69-C71BEAFCDAF8@gmail.com>
Subject: Re: [PHP-DEV] [RFC] clamp
From: hallbergkim@gmail.com (Kim Hallberg)

> On 28 Jun 2021, at 11:16 AM, Nikita Popov <nikita.ppv@gmail.com> =
wrote:
>=20
> The RFC is missing a precise description of how this function works =
with NaN and negative zero. The expected behavior is that if min or max =
are NaN, an exception is thrown, if num if NaN then NaN is returned, and =
the behavior wrt negative zero is unspecified. This makes for an =
IEEE754-2008 style clamping operation. Making negative zero smaller than =
positive zero would be an IEEE754-2019 style clamp.
>=20
> I'm not really convinced that this is a worthwhile addition, but also =
not particularly opposed to it.
>=20
> Regards,
> Nikita

Will update the RFC with a more precise description. Didn=E2=80=99t know =
there existed IEEE standards for clamping operations. Adding exceptions =
for NaN values into min/max is something I agree would be expected =
behaviour since a non number in the range don=E2=80=99t really make =
sense.

As for the negative zero behaviour though, I=E2=80=99m not sure how to =
address that since frankly it=E2=80=99s out of my league, and when would =
a range of -0, 0 make sense to have? Cannot see a value being placed =
between that expect for 0, so a `clamp(0, -0, 0)` would return 0 then?

Thanks, Kim.