Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:79403
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain ajf.me designates 192.64.116.216 as permitted sender)
References: <001201d00f41$cb284810$6178d830$@tutteli.ch>
Mime-Version: 1.0 (1.0)
In-Reply-To: <001201d00f41$cb284810$6178d830$@tutteli.ch>
Content-Type: text/plain;
	charset=us-ascii
Content-Transfer-Encoding: quoted-printable
Message-ID: <958E6385-B585-47DA-A305-11D2D1CF8F6B@ajf.me>
Cc: internals <internals@lists.php.net>
Date: Wed, 3 Dec 2014 21:55:28 +0000
To: Robert Stoll <php@tutteli.ch>
Subject: Re: [PHP-DEV] function overloading
From: ajf@ajf.me (Andrea Faulds)

Hi!

> On 3 Dec 2014, at 21:40, Robert Stoll <php@tutteli.ch> wrote:
>=20
> Heya,=20
>=20
>=20
>=20
> I would like to know If it is somehow possible to overload existing functi=
ons by extensions. And if it is possible, are
> there already some extension doing it?

Possibly in an extension, maybe. But as a weakly-typed, dynamic, reflective l=
anguage, overloading of functions is both a bad idea (unpredictable program e=
xecution) and unnecessary (types can be checked at runtime, optional paramet=
ers exist). I also don't think overloading of functions is a good idea gener=
ally, it leads to horrible APIs where there are 28 different versions of a f=
unction each taking differently-typed arguments in different orders. I am gl=
ad PHP largely lacks this madness, it makes it an easier language to read an=
d to write. Overloading isn't needed: use optional parameters, a named optio=
ns associative array, or check argument types within the function. If what y=
ou're doing doesn't fit into those categories, maybe you need to make a sepa=
rate function.

> I am not talking about the magic __call function. I am talking about somet=
hing like: let's assume GMP has overloaded the
> function "abs" and therefore one can write:
> $n =3D gmp_init(-12345678901234567890);
> abs($n+ 2);  // instead of using gmp_abs
>=20
>=20
>=20
> I know that GMP has overloaded operators so it doesn't seem so far that ov=
erloading functions is possible as well.

Being able to change APIs provided by other libraries from your library soun=
ds like a bad idea.

If you just want abs() to work on bignums, my Big Integer Support RFC would h=
andle that. :)

--
Andrea Faulds
http://ajf.me/=