Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:97841
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.161.177 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <cf10d181-4add-cc79-4dcf-dc5c49479dd7@fleshgrinder.com>
References: <C098D88D-32D3-4489-B1FA-C49527B28979@gmail.com>
 <CAF+90c8jVhaOu_GF9hAR-_VTkm9cQMV86RS09-P_cKtORFrQVg@mail.gmail.com> <cf10d181-4add-cc79-4dcf-dc5c49479dd7@fleshgrinder.com>
Date: Tue, 17 Jan 2017 23:45:27 +0100
Message-ID: <CAF+90c-mQBVzWGooJf+fjJpk-SssrADs_5ey_3w-dHuh_PcU5w@mail.gmail.com>
To: PHP internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary=94eb2c0bb4d47f1d920546520eae
Subject: Re: [PHP-DEV] [RFC][DISCUSSION] - Disallow Multiple Constructor Calls
From: nikita.ppv@gmail.com (Nikita Popov)

--94eb2c0bb4d47f1d920546520eae
Content-Type: text/plain; charset=UTF-8

On Tue, Jan 17, 2017 at 9:52 PM, Fleshgrinder <php@fleshgrinder.com> wrote:

> On 1/17/2017 9:48 PM, Nikita Popov wrote:
> > Without putting a lot of though into this, the restriction sounds
> > reasonable.
> >
> > However, it is not immediately clear to me how this can be implemented
> > without adding a disproportionate amount of overhead for something which
> I
> > would consider a rather minor feature. If an object only had a single
> > constructor this would be a simple flag, but if you want to allow one
> call
> > per each constructor in the hierarchy (i.e. you can call the constructor
> > and each parent constructor once), this information needs to be stored
> > somehow. The obvious way would be a hashset containing the constructors
> (or
> > CEs) that have been called on the object, but I don't think this would be
> > acceptable.
> >
> > Regards,
> > Nikita
> >
>
> Hey Nikic!
>
> Thanks for skimming over it. I also though about implementation already.
> My initial though was -- especially to keep it simple -- to simply
> delete the method from the object's table.
>
> Not sure if this is possible or just introduced more problems than it
> solves. Also considering that we think reflection should continue to
> allow calling it multiple times, though this is obviously still debatable.
>

Methods are stored on the class, with the object containing a pointer to
the class. As such, it's not possible to delete a method from a single
object, you can only delete a method from the entire class. (In practice
you can't do that either ^^)

Nikita

--94eb2c0bb4d47f1d920546520eae--