Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:108009
User-Agent: Cyrus-JMAP/3.1.7-731-g1812a7f-fmstable-20200106v2
Mime-Version: 1.0
Message-ID: <9126e194-fee1-4d54-968c-d964d2e442f3@www.fastmail.com>
In-Reply-To: 
 <CAF+90c91GyvTc6XOAVnp4n15LL-cqFHDMtPA8_QLnvmaiWOiwA@mail.gmail.com>
References: <a68ab6cf-6a26-43c6-9b12-c41366cce01f@www.fastmail.com>
 <CAF+90c91GyvTc6XOAVnp4n15LL-cqFHDMtPA8_QLnvmaiWOiwA@mail.gmail.com>
Date: Mon, 06 Jan 2020 12:17:44 -0600
To: "php internals" <internals@lists.php.net>
Content-Type: text/plain
Subject: =?UTF-8?Q?Re:_[PHP-DEV]_Inconsistent_class_behavior_and_undocumented(=3F?=
 =?UTF-8?Q?)_BC_change?=
From: larry@garfieldtech.com ("Larry Garfield")

On Mon, Jan 6, 2020, at 5:16 AM, Nikita Popov wrote:

> Others have already explained why constructors are exempted from LSP
> checks, so let me reply to your other point...
> 
> I believe that your example should indeed be legal in general and created a
> PR to fix this: https://github.com/php/php-src/pull/5059
> 
> With this change, the variadic argument can replace any number of
> non-variadic arguments, as long as it is compatible with them, where
> compatibility is, as usual, determined based on type contravariance, as
> well as reference passing invariance. Effectively this means that
> function(...$args) is compatible with all signatures that do not involve
> reference passing.
> 
> Does anyone see any soundness issues with this change?
> 
> Nikita

Nifty!  Since `func(...$args)` is already a universal caller, that gives us a symmetrical way to defer calls, either to a parent or to a delegated method on a wrapped object.  So that means one can implement a pass-through class where most of the methods are:

class Wrapper implements Thing {
  public function thing(...$args) { $this->wrapped->thing(...$args); }
}

The symmetry appeals to me, even if it means types get caught only at the lower level.

One possible question; would this mean bypassing type checks if not delegating?  Viz:

interface Foo {
  public function bar(Foo, $f, Baz $b, Beep $e);
}

class Impl implements Foo {

  public function bar(...$args) { ... }

}

$i = new Impl();

// Will this still report a type error, and if so, where/how?
// If not, does that become a back-door way to dodge an interface?  Is that good?
$i->bar('a', 'b', 'c');

--Larry Garfield