Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:122719
Precedence: bulk
MIME-Version: 1.0
References: <CAPzBOBMGe=kWWiJYGLG02qxh+2Pe5yTqodhGtAE3LjdAPmkQUw@mail.gmail.com>
 <CADyq6sKSLXXAGm4snK2tSV8fw3K-36LXKWF0wVJ7ZxwZfL_9PA@mail.gmail.com>
In-Reply-To: <CADyq6sKSLXXAGm4snK2tSV8fw3K-36LXKWF0wVJ7ZxwZfL_9PA@mail.gmail.com>
Date: Thu, 21 Mar 2024 18:39:55 -0400
Message-ID: <CAKuGYLTt2ApUO4VW3zyf=X=7+nHxR2BRjx_s55X3S84Eg+vzkQ@mail.gmail.com>
Subject: Re: [PHP-DEV] Proposal: AS assertions
To: Marco Pivetta <ocramius@gmail.com>
Cc: Robert Landers <landers.robert@gmail.com>, internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary="000000000000960e88061433635d"
From: matthewmatthew@gmail.com (Matthew Brown)

--000000000000960e88061433635d
Content-Type: text/plain; charset="UTF-8"

>
> What's the advantage of a language construct over the following?
>
> ```php
> /**
>  * @template T of object
>  * @psalm-assert T $value
>  * @param class-string<T> $type
>  */
> function as(mixed $value, string $type): mixed
> {
>     if (! $value instanceof $type) { throw
> SomeKindOfException::forMismatchingRequirements($value, $type); }
>
>     return $value;
> }
>
> echo as(myExpression(), MyType::class)->methodOfMyType();
> ```
>
>
A static analysis tool supporting something in docblocks should not
preclude those things being added in syntax IMO.

Hack's `as` operator is very neat, and it'd be trivial for existing SA
tools to support the same in PHP.

Here's a demo of `as` in Hack getting flagged by a SA tool:
https://hakana.dev/#XQAAgAA2AAAAAAAAAAAzHUn_qWH7EwabJzyN0tdfxv3ug6f7oZ-qScnamcl1qjUZCPmuKA3tD-KFr1f0ZPcrAXt_D1L___KsQAA%3D

This would also benefit from a `nonnull` subtype of `mixed` which could be
used as a null refinement.

--000000000000960e88061433635d
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div class=3D"gmail_quote"><blockquote class=3D"gmail_quot=
e" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204)=
;padding-left:1ex"><div dir=3D"ltr"><div class=3D"gmail_quote">What&#39;s t=
he advantage of a language construct over the following?<br><div><br></div>=
<div>```php</div><div>/**</div><div>=C2=A0* @template T of object<br></div>=
<div>=C2=A0* @psalm-assert T $value</div><div>=C2=A0* @param class-string&l=
t;T&gt; $type</div><div>=C2=A0*/<br></div><div>function as(mixed $value, st=
ring $type): mixed</div><div>{</div><div>=C2=A0=C2=A0=C2=A0 if (! $value in=
stanceof $type) { throw SomeKindOfException::forMismatchingRequirements($va=
lue, $type); }<br></div><div><br></div><div>=C2=A0 =C2=A0 return $value;</d=
iv><div>}</div><div><br></div><div>echo as(myExpression(), MyType::class)-&=
gt;methodOfMyType();</div><div>```</div><div><br></div></div></div></blockq=
uote><div><br></div><div>A static analysis tool supporting something in doc=
blocks should not preclude those things being added in syntax IMO.<br><br>H=
ack&#39;s `as` operator is very neat, and it&#39;d be trivial for existing =
SA tools to support the same in PHP.</div><div><br></div><div>Here&#39;s a =
demo of `as` in Hack getting flagged by a SA tool:=C2=A0<a href=3D"https://=
hakana.dev/#XQAAgAA2AAAAAAAAAAAzHUn_qWH7EwabJzyN0tdfxv3ug6f7oZ-qScnamcl1qjU=
ZCPmuKA3tD-KFr1f0ZPcrAXt_D1L___KsQAA%3D">https://hakana.dev/#XQAAgAA2AAAAAA=
AAAAAzHUn_qWH7EwabJzyN0tdfxv3ug6f7oZ-qScnamcl1qjUZCPmuKA3tD-KFr1f0ZPcrAXt_D=
1L___KsQAA%3D</a><br><br>This would also benefit from a `nonnull` subtype o=
f `mixed` which could be used as a null refinement.</div></div></div>

--000000000000960e88061433635d--