Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:92597
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 74.125.82.45 as permitted sender)
To: "guilhermeblanco@gmail.com" <guilhermeblanco@gmail.com>,
 Jesse Schalken <me@jesseschalken.com>
References: <CADqTB_gUbfu91QRm1U0KyDjai8DKw7yCE6QxgUecy60oq3Zi8Q@mail.gmail.com>
 <CAKiSzdCO83XM0zqAeFnjQEMJvbd0yBmYAFnj0H6AtDumofVbkw@mail.gmail.com>
 <CAKiSzdCCEF=7OWtvZETG3bRoByCCLqEAEjuLW_H6m7MmDtJ8_g@mail.gmail.com>
 <CAFKaYMoQUUFLqgDjOHj3W0+a63u3s=vqnU5gWMCj0zvmvXFF5Q@mail.gmail.com>
 <3cc8a4c7-2640-11ae-a67b-06f909ac1e27@texthtml.net>
 <CAESVnVqCZnrbHa+5_tSYG70VyvGfpZoDM98bEst3pg36HahYEg@mail.gmail.com>
 <57173859.4080501@rochette.cc>
 <CAFKaYMq_KFrygXwcuTd85y11X0_YMFop2tum+8jjXnkGKON6QA@mail.gmail.com>
 <CAFp73Xu2dM+E9Ri8dN-yt9Ha6-1UmOtD0RYq+HZXbjvzcWqJwg@mail.gmail.com>
 <CAFKaYMpWN+9uBqJ_p6-p14ZqYm0cm+q3LHkY3o9q7EvFCBS9Bg@mail.gmail.com>
 <CAFKaYMo5rtXeV+uE85yuQ_xOpREiKF-zVdQN6T-vDET8WhFvzg@mail.gmail.com>
 <CAF1yvAQh2O6gdczOa6X5TTi55Nd8AMrjjmR9LX+szmy8aaHdxA@mail.gmail.com>
 <57180921.8020000@gmail.com>
 <CAF1yvASAJoBp86Yr8gNq0KmgrOm21a0vC5vMQ8s2UszXBoVEJg@mail.gmail.com>
 <CAFp73Xv7d2vpoCtTeghp5LCfQOjMLtLA=6-D7p5-vJhw+PLnEg@mail.gmail.com>
Cc: PHP internals <internals@lists.php.net>
Message-ID: <5718F530.1070503@gmail.com>
Date: Thu, 21 Apr 2016 16:43:44 +0100
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:38.0) Gecko/20100101
 Thunderbird/38.7.2
MIME-Version: 1.0
In-Reply-To: <CAFp73Xv7d2vpoCtTeghp5LCfQOjMLtLA=6-D7p5-vJhw+PLnEg@mail.gmail.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] [RFC:generics]
From: rowan.collins@gmail.com (Rowan Collins)

guilhermeblanco@gmail.com wrote on 21/04/2016 16:25:
> I understand others mentioned we could create a new interface that 
> wraps the other two interfaces, but that still doesn't solve the 
> problem where I want to extend a class AND also implement a specific 
> interface.

I covered this case in my previous mail: you can create an abstract 
class that extends a concrete class and adds an interface to it, and the 
extra methods are implicitly added as abstract for descendants to implement:

// This could be somewhere else in code you don't want to touch
class Foo { public function doStuff() { /* ... */ } }
interface Bar { public function neededAction(); }

// Wherever you need to use the intersection Foo & Bar
abstract class FooBar extends Foo implements Bar {}
class Foo<T instanceof FooBar> { /* can rely on both T#doStuff and 
T#neededAction existing */ }

// The implementation of the intersection must be in our control, to use 
"FooBar" rather than "Foo" as a base
// This is also true of extending interfaces, since PHP doesn't do 
duck-typing of interfaces
class FooBarImplementation extends FooBar { public function 
neededAction() { /* ... */ } }


> What happens if the initial one get approved but it is too late to 
> propose the subsequent one? Then PHP 7.1 would be released with a 
> partial implementation and not fully featured as it should.

No other part of the language allows such type algebra, so a generics 
implementation that doesn't allow it wouldn't feel "incomplete" to 
anyone who wasn't aware of the discussion. Do you consider the typehints 
PHP has had for over a decade to be "a partial implementation" because 
they don't include this feature?

I'm assuming that if generics passed, and type algebra passed, then 
generics-with-type-algebra would go through on the nod and make it to 
the same version.

Regards,
-- 
Rowan Collins
[IMSoP]