Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:14388
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Reply-To: lpedretti@suserver.com
Organization: Axon Sistemas
To: internals@lists.php.net
Date: Fri, 14 Jan 2005 13:01:43 -0300
User-Agent: KMail/1.7
References: <20050110140943.99185.qmail@web41415.mail.yahoo.com> <200501101315.32968.lpedretti@suserver.com> <Pine.LNX.4.58.0501102157300.27172@localhost>
In-Reply-To: <Pine.LNX.4.58.0501102157300.27172@localhost>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Message-ID: <200501141301.43162.lpedretti@suserver.com>
Subject: Re: [PHP-DEV] Basic types hinting and variable hinting
From: lpedretti@suserver.com (Leonardo Pedretti)



On Monday 10 January 2005 17:58, Derick Rethans wrote:
> On Mon, 10 Jan 2005, Leonardo Pedretti wrote:
> > Are there any plans to include type hinting with basic types and type
> > hinting for variables? (such as optional declaration)
>
> There are no such things planned, except for "array" afaik.
That is quite good already, but...
> > It would be useful to declare the type of a data member of a class, for
> > example, for automation of data fetching from SQL databases or such
> > things.
>
> That is *definitely* not on the list - PHP is weakly typed and will
> remain so.
I can't understand why being able to point some variable as strongly typed =
and=20
remaining the default as weak typed would make php strongly typed. In fact,=
=20
that assumption IS incorrect, a language is weakly typed when variable type=
s=20
are not forced to be respected though they already exist and change=20
dynamically, and the programmer actively having to point a variable as=20
strongly typed for the scope of a function obviously does not contradict th=
at=20
postulate.
Even more, just having that feature in object data members would be heaven.=
=20
Basic type hinting can be emulated using is_int, is_string, and so on on th=
e=20
beginning of the function needing that, but for objects it can not be done =
to=20
know the type of a data member previously to it's assignment, and having to=
=20
assign a value in the constructor forces one to declare a specific=20
constructor doing that for classes that derive or implement an interface th=
at=20
has functions using that type for automation of tasks. Code just would be=20
really so cleaner just by allowing one to put a type in front of an object=
=20
data member declaration and allowing to retrieve that type.

=46or example, consider this code:

abstract class base {
 function __construct() {
  // constructor just sets up some things
 }
 function someoperation() {
  /* this function needs to load the data members of basic types with some=
=20
clear values but everything that is an object or an array must be filled=20
through some specific process that can be automated just knowing the specif=
ic=20
type in advance */
 }
}

class specific_class {
 function construct() {
  $var1 =3D new SomeClass;
  $var2 =3D new SomeOtherClass;
 }
 private $var1;
 private $var2;
}

would the types could be showed int the class, function someoperation could=
=20
still perform its task without any change, but specific_class now would be=
=20
just the following:

class specific_class {
 private SomeClass $var1;
 private SomeOtherClass $var2;
}

Which is identical to the upper version. Moreover, specifying the construct=
or=20
parameters in the declaration should not be a problem, as it is equivalent=
=20
code to the upper version.

Well, i hope all this can be useful!

Regards
Leo

>
> Derick
>
> --
> Derick Rethans
> http://derickrethans.nl | http://ez.no | http://xdebug.org

=2D-=20

Leonardo Pedretti
Axon Sistemas
L=EDder de Equipo Proyecto Basalto