Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:47892
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain openmv.com from 64.15.152.204 cause and error)
To: "'Stefan Marr'" <php@stefan-marr.de>,
	"'Lukas Kahwe Smith'" <mls@pooteeweet.org>
Cc: "'Derick Rethans'" <derick@derickrethans.nl>,
	"'PHP Developers Mailing List'" <internals@lists.php.net>
References: <alpine.DEB.2.00.1004120931260.8571@kossu.derickrethans.nl> <03128BF8-9EAC-4BAC-AAFE-6D8AA21153DC@pooteeweet.org> <4DA7D942-6BF1-4A0F-AECF-3EB82900798D@stefan-marr.de>
In-Reply-To: <4DA7D942-6BF1-4A0F-AECF-3EB82900798D@stefan-marr.de>
Date: Mon, 12 Apr 2010 10:11:51 -0400
Message-ID: <003701cada4a$196d29a0$4c477ce0$@com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
thread-index: AcraIQNxVP66ijEoTIWvbthLzNeoPQAJ3mow
Content-Language: en-ca
Subject: RE: [PHP-DEV] Traits
From: jbondc@openmv.com ("Jonathan Bond-Caron")

On Mon Apr 12 05:16 AM, Stefan Marr wrote:
> 
> On 12 Apr 2010, at 10:39, Lukas Kahwe Smith wrote:
> > On 12.04.2010, at 10:34, Derick Rethans wrote:
> >> Hi!
> >>
> But just as a quick response, without aliasing, there would be no way 
> to use a conflicting method from a trait.
> On first sight, that is not a really good thing.

Hi Stefan, is it possible to have renaming and aliasing? 

Some context, say we have: 

trait Something { 
  function renameIssue() {
	$method = 'call';
      $this->$method();

	call_user_func(array($this, $method));
  }
  function call () {
     echo "a";
  }
}

class Foo {
   use Something {
     function call() as call2();
   }
}

With renaming:

$f = new Foo;
$f->call2(); 		// a
$f->renameIssue();   	// call() method does not exist

With aliasing:

$f = new Foo;
$f->call(); 		// a
$f->call2(); 		// a
$f->renameIssue();   	// a

While aliasing prevents the error, the syntax could re-use the 'clone'
keyword:

class Foo {
   use Something {
     function call() clone call2();
   }
}

So we'd have both behaviors:
class Foo {
   use Something {
     function call() as call2();
     function call() clone call3();
   }
}

We have call(), call2(), call3()

class Foo {
   use Something {
     function call() as call2();
   }
}

We have call2()

Other than that, the traits proposal still feels incomplete compared to
grafts since from Stefan's research: "the lack of state means that virtually
all traits are incomplete"

It's still a great step if the current proposal (methods only) gets
committed.