Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:93549
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.192.173 as permitted sender)
To: Bob Weinand <bobwei9@hotmail.com>
References: <CAL=_i_mcJoFnPgeU5FQCJN=SDYnNK7Jz8zoVMv9EMuWaGRYSMg@mail.gmail.com>
 <BY2PR0201MB17845690B6386D75AF4C3D79BF4F0@BY2PR0201MB1784.namprd02.prod.outlook.com>
 <CAL=_i_=EqNoHvXmY+RdQW-reC4YF8AkuWFtGGJGOOSqX4a010g@mail.gmail.com>
 <cfe8404a-54f6-24b7-0ba8-8f807c820821@zend.com>
 <CAL=_i_kze8MNSmpg4WT+EJCRPmvty=8DA6DkuHSx+WCBzJjytg@mail.gmail.com>
 <CAF+90c9cBqKfy_q5rME1u56Rq366niYUBDTY+3_2iFJSZD9fDA@mail.gmail.com>
 <CANUQDCj0O6aVwTqo+CcAbX8AWvwPQHfhRtsYRUrMS_aFjHM5RQ@mail.gmail.com>
 <7B.12.14311.F79C5475@pb1.pair.com>
 <1b12b09f-f190-dca0-51d9-468e9c571268@fleshgrinder.com>
 <20160525213546.D7EE41A801B3@dd1730.kasserver.com>
 <ff892ae1-6dc9-cf2c-22a5-5f13a883efdc@gmail.com>
 <BLU436-SMTP2139668C03C06570F6D0BD2E2410@phx.gbl>
Cc: Thomas Bley <mails@thomasbley.de>, ajf@ajf.me, internals@lists.php.net,
 ocramius@gmail.com, bensor987@neuf.fr, Niklas Keller <me@kelunik.com>
Message-ID: <53274ca9-98a5-ca46-94ac-57b2c61885d4@gmail.com>
Date: Wed, 25 May 2016 22:00:22 -0700
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:45.0)
 Gecko/20100101 Thunderbird/45.1.0
MIME-Version: 1.0
In-Reply-To: <BLU436-SMTP2139668C03C06570F6D0BD2E2410@phx.gbl>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] [RFC][Vote] Typed Properties
From: smalyshev@gmail.com (Stanislav Malyshev)

Hi!

> Also, property access never was safe. Consider:
> 
> $obj = new class {
>     public $foo;
>     function __construct() {
>         unset($this->foo);
>     }
>     function __get($prop) {
>         throw new Exception("No $prop for you today!");
>     }
> };
> var_dump($obj->foo); // exception.

This is different. Your code throws exception purposefully so you intend
something to break here - you might as well just written exit(1) and
said there's no safe code at all since you could exit at any moment.
That's not what I mean. I mean that operation that was previously not
prone to fatal errors just because of variable holding some value would
now produce such errors - not because the user wrote code to
specifically do that but because of hidden properties of the engine. If
the engine would by itself insert such code as above into user classes
that would be dangerous too but of course it does not.

> 
> isset() checks will continue to work. 

For some definition of "work" - since it would return false for null
values, but null values are actually OK.

> It's just about getting a *value*. (and to check whether it's a typed property, there's Reflection support in the RFC.)

You don't propose using Reflection before each variable access I
presume. Reflection is useless in runtime scenario.

-- 
Stas Malyshev
smalyshev@gmail.com