Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:120576
To: PHP Internals List <internals@lists.php.net>
Message-ID: <648A1621.1070306@adviesenzo.nl>
Date: Wed, 14 Jun 2023 21:33:53 +0200
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:38.0) Gecko/20100101
 Thunderbird/38.7.0
MIME-Version: 1.0
Content-Type: multipart/alternative;
 boundary="------------010903050108050109040507"
Subject: Unexpected behaviour when using Traits with final methods
From: php-internals_nospam@adviesenzo.nl (Juliette Reinders Folmer)

--------------010903050108050109040507
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi all,

I'm running into something which peaked my curiousity due to its 
unexpected behaviour, so I'm writing to the list in the hopes of finding 
out whether this is by design, a bug or an oversight which should be 
fixed (via an RFC?).

 > The precedence order is that members from the current class override 
Trait methods, which in turn override inherited methods.
Source: 
https://www.php.net/manual/en/language.oop5.traits.php#language.oop5.traits.precedence

If one would want to prevent the "current class" from being able to 
override a trait method, I would expect that the `final` keyword would 
be able to do this, just like it would when a method is inherited from a 
parent class.

Unfortunately, that does not seem to work and it appears that the 
current class can freely override final methods from a trait:

```
trait Foo {
     final protected function bar() {
         echo 'Foo';
     }
}

class FooBar {
     use Foo;

     public function __construct() {
         $this->bar();
     }

     protected function bar() {
         echo 'FooBar';
     }
}

new FooBar(); // Prints "FooBar".
```
https://3v4l.org/hlirb

Which brings me to my second point: in PHP 8.0 the use of the `final` 
keyword with `private` methods has been turned into a `warning` via this 
RFC: https://wiki.php.net/rfc/inheritance_private_methods

... but the "current class" has access to `private` methods inherited 
from a trait, so I would expect `final private` methods in traits to be 
exempt from that warning, but turns out that that will still throw the 
warning: https://3v4l.org/Vb0S2

Is there anyone who can shed some light on this behaviour ?

Smile,
Juliette

--------------010903050108050109040507--