Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:112320
MIME-Version: 1.0
References: <CAEsg9X0fc-Nq1_U3+8--brZJ031iQnNwi1oL3qOXt-6vERErtA@mail.gmail.com>
 <265e9893-f82c-39f2-3323-2854b246658d@processus.org>
In-Reply-To: <265e9893-f82c-39f2-3323-2854b246658d@processus.org>
Date: Mon, 30 Nov 2020 16:19:01 +0100
Message-ID: <CAF+90c_J2oOkjJ0xUup5hbdupe7K6beqiFyiqDxG-_pzkzoy6A@mail.gmail.com>
To: "Pierre R." <pierre-php@processus.org>
Cc: David Rodrigues <david.proweb@gmail.com>, PHP Internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary="000000000000a7b82005b5548820"
Subject: Re: [PHP-DEV] Traits "implements" support
From: nikita.ppv@gmail.com (Nikita Popov)

--000000000000a7b82005b5548820
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Thu, Nov 26, 2020 at 5:53 PM Pierre R. <pierre-php@processus.org> wrote:

> Le 26/11/2020 =C3=A0 17:50, David Rodrigues a =C3=A9crit :
> > Hello!
> >
> > Now with PHP 8, traits supports abstract functions that should be
> > implemented by the class that uses this trait.
> >
> > So is it now possible that traits could have support to implements too?
> So
> > when class uses this trait, it will be an implementation of X interface
> and
> > will have to implement methods based on it.
> >
> > interface A {
> >      public function x();
> > }
> >
> > trait B implements A {
> >      public function y() { $this->x(); }
> >      // public function x() still not implemented.
> > }
> >
> > class Z { // impliticy implements A interface due to B trait.
> >      use B;
> >      public function x() {} // required!
> > }
> >
> >
> > Atenciosamente,
> > David Rodrigues
> >
> Hello,
>
> I'd love this too, but I'm not sure what would be the static code
> validation behavior, knowing that you can rename methods of traits when
> using them in a class: should the trait method not be validated ? Should
> the final class be validated after the trait has been used, meaning that
> if a method name change, the final class must re-implement it ?
>

Common misconception: It's not possible to rename methods of traits, you
can only create aliases. That said, there's still an issue in that you
might resolve a conflict in favor of an identically named method that uses
a different signature.

Note that Hack uses a "require implements" syntax instead, which makes it
clear that the interface needs to be implemented by the using class:
https://docs.hhvm.com/hack/classes/trait-and-interface-requirements

Regards,
Nikita

--000000000000a7b82005b5548820--