Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:96929
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 74.125.82.47 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <CAEsg9X1uBMuTubtuo0ULkhPLr22HaNSFBrijnw7S7TO7Wuxhgg@mail.gmail.com>
References: <CAA69SC4th-Hkxj162xn3RWAG44tFkhGTo3Pm=mephPdSHfcHsw@mail.gmail.com>
 <CAEsg9X1uBMuTubtuo0ULkhPLr22HaNSFBrijnw7S7TO7Wuxhgg@mail.gmail.com>
Date: Wed, 16 Nov 2016 16:47:08 +0100
Message-ID: <CABdc3WpsOrV1B5BAvf_c-_RhnnWHSUCWwJztcL4Mo7N-DhZ-cw@mail.gmail.com>
To: David Rodrigues <david.proweb@gmail.com>
Cc: PHP Internals List <internals@lists.php.net>, =?UTF-8?Q?Silvio_Mariji=C4=87?= <marijic.silvio@gmail.com>
Content-Type: multipart/alternative; boundary=089e01493e645dcf5305416cfcf8
Subject: Re: [PHP-DEV] Immutability RFC
From: michal.brzuchalski@gmail.com (=?UTF-8?Q?Micha=C5=82_Brzuchalski?=)

--089e01493e645dcf5305416cfcf8
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

16.11.2016 16:39 "David Rodrigues" <david.proweb@gmail.com> napisa=C5=82(a)=
:
>
> 2016-11-16 11:57 GMT-02:00 Silvio Mariji=C4=87 <marijic.silvio@gmail.com>=
:
> > If we treat those objects as values then this should return true. But
then
> > again there might be some confusion because then two operators are doin=
g
> > the same thing. Maybe throw an error ? Suggestions ?
>
> I just thinking about "how it will be done"?
> I imagine that we should store all calling to memory, then I see two
> problems for this case:
>
> 1. High memory consumption: all immutable objects should be stored on
> memory eternally (or until the script ends). I mean: if I call it a
> hundred times with differents parameters, I'll have a hundred of used
> memory, that will avoid any used variables of be GC (once that it
> should stores it for references).
>

All instances are GC the same as nornal classes when unneeded any longer.

> 2. High CPU consumption: every time that I instantiate an immutable
> class, the engine will loop on each possibilities and check if all
> parameters match to then return the same object (or then, create a
> new). Then the major question for it is: should all parameters of an
> immutable constructor be considered on instantiate? Should I pass the
> exactly copy (read "reference") of object to the parameter to return
> the same instance or a "shallow copy"?
>

There is no cache - each immutable objects is created the same as normal
object except it won't let you pass references.

> For both problems above, I guess that a "factory pattern with cache"
> should works better, once that I could consider the exact value where
> it requires to be immutable, and I could cache only if it is really
> required, for instance:
>
> class SomeFactory {
>     private static $cache =3D [];
>
>     public static function getType(Type $type) {
>         if (!$type->shouldBeCache()) {
>             // Code flexibility...
>             return new self($type);
>         }
>
>         if (!array_key_exists($type->id, self::$cache) {
>             self::$cache[$type->id] =3D new self($type);
>         }
>
>         return self::$cache[$type->id];
>     }
> }
>
>
> --
> David Rodrigues
>
> --
> PHP Internals - PHP Runtime Development Mailing List
> To unsubscribe, visit: http://www.php.net/unsub.php
>

--089e01493e645dcf5305416cfcf8--