Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:98094
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain hotmail.com designates 65.55.111.176 as permitted sender)
Message-ID: <BLU436-SMTP150D957E1AB6FB236308F2EE24A0@phx.gbl>
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0 (Mac OS X Mail 10.1 \(3251\))
In-Reply-To: <D1.1B.51557.257F0985@pb1.pair.com>
Date: Tue, 31 Jan 2017 22:18:45 +0100
CC: internals@lists.php.net
Content-Transfer-Encoding: quoted-printable
References: <CAFMT4Nrt7XPMnELoAGLu=sSRCTCtzPCTaSE8HJjKfHjDXOXU_A@mail.gmail.com>
 <E6.ED.51557.B64EF885@pb1.pair.com>
 <3F428CA4-8211-44E6-9B60-62ADB47934B3@koalephant.com>
 <5bfde844-beee-0f15-1f92-f77607f70ae6@mabe.berlin>
 <D1.1B.51557.257F0985@pb1.pair.com>
To: Andrea Faulds <ajf@ajf.me>
Subject: Re: [PHP-DEV] [RFC][Discuss] Arrow Functions
From: bobwei9@hotmail.com (Bob Weinand)


> Am 31.01.2017 um 21:45 schrieb Andrea Faulds <ajf@ajf.me>:
>=20
> Hi Marc,
>=20
> Marc Bennewitz wrote:
>> - Also I like the "use"  keyword you have to define your variables
>>=20
>>    -> Would it be helpful to allow "function () use (*) {}" to inline
>> all available variables?
>=20
> I did think of that, but it's not as concise as not having to specify =
`use` at all. You could take the opposite approach and auto-capture by =
default, but permit `use ()` to capture nothing, alongside regular `use =
(...)` for explicit capture of variables.
>=20
> Thanks.
>=20
> --=20
> Andrea Faulds
> https://ajf.me/

Hey Andrea,
you realize that you are actually proposing quite a heavy BC break?

Any code which uses the same variable than in parent scope and does not =
overwrite it before its first usage will break.

E.g.:

$foo =3D [1,2,3];
// more code working on foo
doSomeAction(function($bar) {
    // waaah, $foo is now [1,2,3] instead of null
    foreach ($bar as $x) {
        $foo[] =3D $x;
    }
    return $foo;
});

Or what I have seen in real code:

$response_text =3D "Do action";
store_response($response_text);
// much later
doSomeAction(function() {
    $response_text .=3D "Do other action"; // Note that accidental .=3D =
instead of =3D
    store_response($response_text);
});

The code still works as expected, because things are NULL-initialized =
currently.
Now with implicit auto-capture this is breaking. No ahead of time errors =
like parse or compile errors, which warn you. Just subtly breaking. =
Depending on where it's done, it may not be even noticed soon (variables =
with the same name at different places tend to have compatible types), =
just misbehave subtly.

This is not acceptable.

Now, when you suggest "function($bar) use (*) =3D> $foo + $bar", I feel =
like half the point of the RFC, the conciseness of the syntax, is =
missed.
And I'd also reject function($bar) =3D> $foo + $bar as this would mean =
the "function" keyword introducing different semantics depending on what =
follows. [apart from being still not as concise as "fn"]

Thanks,
Bob=