Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:109029
To: "internals@lists.php.net" <internals@lists.php.net>
Thread-Topic: RFC idea: Block scoped variables with "let $x = expr"
Thread-Index: AQHV+ug0CGaQ9aFqj0aNq48f2k+kqQ==
Date: Sun, 15 Mar 2020 16:47:30 +0000
Message-ID:
 <DM5PR07MB3067D2DEAD4B3C85FBDFD559F9F80@DM5PR07MB3067.namprd07.prod.outlook.com>
Accept-Language: en-CA, en-US
Content-Language: en-CA
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-MS-Exchange-CrossTenant-RMS-PersistedConsumerOrg: 00000000-0000-0000-0000-000000000000
X-MS-Exchange-CrossTenant-Network-Message-Id: 69b74b9a-9699-4798-a9da-08d7c9008d9e
X-MS-Exchange-CrossTenant-rms-persistedconsumerorg: 00000000-0000-0000-0000-000000000000
X-MS-Exchange-CrossTenant-originalarrivaltime: 15 Mar 2020 16:47:30.8213
 (UTC)
X-MS-Exchange-CrossTenant-fromentityheader: Internet
X-MS-Exchange-CrossTenant-id: 84df9e7f-e9f6-40af-b435-aaaaaaaaaaaa
X-MS-Exchange-Transport-CrossTenantHeadersStamped: DM6NAM11HT191
Subject: RFC idea: Block scoped variables with "let $x = expr"
From: tysonandre775@hotmail.com (tyson andre)

Hi internals,=0A=
=0A=
In PHP, variables are currently scoped to the function scope, and can't be =
scoped to a block scope. This makes it difficult to reason about how a vari=
able will actually be used at a glance, especially in long functions, or to=
p-level statement lists of a file.=0A=
(or how the variable was originally intended to be used)=0A=
=0A=
The function scope lifetime of variables in PHP is similar to how JavaScrip=
t treated variables with `var`, before the introduction of the `let` statem=
ent in JS:=0A=
https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Statement=
s/let=0A=
=0A=
Would there be any interest in adding something similar to this in PHP?=0A=
(Or have similar ideas been proposed in the past? I didn't see any.)=0A=
Correctness would be enforced as a best-effort at compile-time - the variab=
les would continue to only be freed at the end of the function call.=0A=
(e.g. __destruct() only gets called when the variable gets reassigned or th=
e function call ends)=0A=
=0A=
- Freeing it immediately would likely require the equivalent of a try{} fin=
ally{} to handle exceptions.=0A=
   try statements prevent opcache from optimizing a function, the last time=
 I checked.=0A=
=0A=
```=0A=
{=0A=
   let $myObject =3D new MyClass(), $flag =3D true;=0A=
   $myObject->process($flag);=0A=
}=0A=
// Accessing or setting $myObject or $flag outside a different let is an E_=
COMPILE_ERROR=0A=
// because it was declared anywhere else in the function body=0A=
// as a let statement=0A=
$myObject->doSomething();=0A=
=0A=
// E_COMPILE_ERROR to access $key or $value after, outside of a separate `l=
et` scope=0A=
foreach ($arr as let $key =3D> let $value) {=0A=
    // It would be a compile error to declare a gotoLabel: inside of a scop=
e of a let statement.=0A=
}=0A=
// E_COMPILE_ERROR to access $key or $value after=0A=
=0A=
echo $$var;  // E_COMPILE_ERROR to use variable variables in the same scope=
, to enforce this as much as possible.=0A=
=0A=
{=0A=
   let $outer =3D 1;=0A=
   {=0A=
       let $outer =3D 2;  // E_COMPILE_ERROR to declare in a different scop=
e, can only reassign=0A=
       $outer =3D 3;   // this is allowed=0A=
    }=0A=
}=0A=
=0A=
{=0A=
    let $myRef =3D&$other;=0A=
}=0A=
=0A=
{=0A=
    let $myRef =3D true;  // This removes myRef from the reference group be=
fore assigning a new value to myRef - $other is unmodified.=0A=
}=0A=
```=0A=
=0A=
I say "best-effort" because require() and $GLOBALS and get_defined_variable=
s() can still read or modify or make references to variables outside of a `=
let`,=0A=
depending on what's in the function body or top-level statements.=0A=
=0A=
- Tyson=