Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:127808
Content-Type: multipart/signed;
	boundary="Apple-Mail=_CE91984A-912F-47A0-BC55-ED2CFEAABF0D";
	protocol="application/pgp-signature";
	micalg=pgp-sha256
Precedence: bulk
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3826.600.51.1.1\))
Subject: =?utf-8?Q?Re=3A_=5BPHP-DEV=5D_Request_for_RFC_Karma_=E2=80=93_kha?=
 =?utf-8?Q?ledalamxyz?=
Date: Mon, 30 Jun 2025 11:02:37 -0500
References: <CAJS5taDf3Sg-phPNN2z7HGOuV5WxMKJf9vuyj07kKrUKUonD0g@mail.gmail.com>
 <CAPyj-LAhtmPHG3P+sHVSzKLBT95Y2Vx2YANoBy4pp3bt8xSMcw@mail.gmail.com>
 <CAJS5taCHvmHGxixuaNYZ8kTYhHXTom4AsKfGNFNDypBsnJ0j2A@mail.gmail.com>
To: PHP internals <internals@lists.php.net>
In-Reply-To: <CAJS5taCHvmHGxixuaNYZ8kTYhHXTom4AsKfGNFNDypBsnJ0j2A@mail.gmail.com>
Message-ID: <92E75FAC-70EA-46C1-AAFC-CC423A2BE77F@php.net>
From: ramsey@php.net (Ben Ramsey)


--Apple-Mail=_CE91984A-912F-47A0-BC55-ED2CFEAABF0D
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8

> On Jun 30, 2025, at 09:52, Khaled Alam <khaledalam.net@gmail.com> =
wrote:
>=20
> Hi Ilija, Ben, and Thomas,
>=20
> Thanks for the rapid feedback. You=E2=80=99re right that PHP=E2=80=99s =
existing expression-throw idioms cover some use cases, but `=3D>!` fills =
three gaps:
> 1. Comprehensive =E2=80=9Cfalsy=E2=80=9D check =3D>! consistently =
treats all PHP-falsy values (null, false, 0, "", []) as failure, in a =
single, unambiguous operator.// Existing: must choose between ??, ?:, =
||, each with its own quirks
> $name ?? throw=E2=80=A6 // only null
> $name ?: throw=E2=80=A6 // also empty strings and "0"
> $name || throw=E2=80=A6 // can=E2=80=99t embed in function call
>=20
> // With `=3D>!`, you get exactly =E2=80=9Cfail on any falsy=E2=80=9D:
> $name =3D>! throw new Exception("Missing name");


You mention =E2=80=9Calso empty strings and `"0"`=E2=80=9D as quirks of =
?:, but the argument you=E2=80=99re making is that `=3D>!` fails on =
=E2=80=9Cany falsy.=E2=80=9D Empty strings and `"0"` are part of =E2=80=9C=
any falsy,=E2=80=9D so I=E2=80=99m a little confused about the =
distinction you=E2=80=99re making here.


> 2. Value forwarding
> All of the above patterns discard or ignore the original value when =
used inline. With =3D>!, the guarded value is both asserted and =
returned, so you can write:
> // Inline validation + use in one expression
> sendEmail( getUserEmail() =3D>! throw new Exception("Email required") =
);
> This both throws on a missing/falsy email and =E2=80=94 when valid =E2=80=
=94 passes the exact email string through to sendEmail().


I don=E2=80=99t understand the distinction you=E2=80=99re making here, =
either. This seems to work currently with ?:

See here: https://3v4l.org/YMLT1#v8.4.8

```
function getUserEmail(): ?string { return 'foo@example.com'; }
function sendEMail(string $email) { echo "Email address is $email"; }

sendEmail( getUserEmail() ?: throw new Exception("Email required") );
```

Unless you mean it also *returns* the falsy value, which in this example =
doesn=E2=80=99t have much value, since it throws an exception, =
disrupting program flow.


> 3. Clean, declarative syntax
> Because it=E2=80=99s a dedicated operator rather than a hacky boolean =
trick, =3D>!:
>     =E2=80=A2 Reads as =E2=80=9Ctake X, or fail=E2=80=9D in one =
glance.
>     =E2=80=A2 Avoids parentheses gymnastics or confusing operator =
precedence.
>     =E2=80=A2 Encourages fluent, lets you chain guards:
> $userId =3D $input["id"] =3D>! throw new Exception("ID missing")
> =3D>! isValidId(...) =3D>! throw new Exception("Invalid ID");


You=E2=80=99ll need to explain the =E2=80=9Cfluent=E2=80=9D behavior in =
more detail. The way I read this is: if `$input["id"]` is falsy, throw =
an exception for the missing ID. If it=E2=80=99s not falsy, however, it =
should proceed to the `isValidId(...)` call, passing it the value. So, =
if the left hand-side is non-falsy, it skips over the immediate =
right-hand side?

If it skips over the immediate right hand side when the value on the =
left is non-falsy, then when it encounters the next `=3D>!`, the =
left-hand side is still non-falsy, so it shouldn=E2=80=99t attempt to =
call `isValidId(...)`, right?


> Demonstration:<image.png>
>=20
> This was primarily an exercise to familiarize myself with the =
internals process. I=E2=80=99d love to gather feedback - and some =
RFC-karma - before drafting a full proposal on the mailing list.
> Thanks again for your time and guidance!


Keep it up! Don=E2=80=99t let our criticism discourage you. :-)

A couple of pointers:

1. Don=E2=80=99t =E2=80=9Ctop post=E2=80=9D on replies to the mailing =
list.
2. It=E2=80=99s generally not a good idea to include images attached to =
mailing list messages.

Cheers,
Ben



--Apple-Mail=_CE91984A-912F-47A0-BC55-ED2CFEAABF0D
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
	filename=signature.asc
Content-Type: application/pgp-signature;
	name=signature.asc
Content-Description: Message signed with OpenPGP

-----BEGIN PGP SIGNATURE-----

iHUEAREIAB0WIQToXQMR3fpbrPOmEOewLZeYnIwHGwUCaGK1HQAKCRCwLZeYnIwH
G9xvAP4mNVttEiaQC7w0FJbOBWp6TOSx4qOnIMs7UZ8pca+00wD+OUx9YdUmn9b9
sQ9K7IpJep7uF8UZFnRFoF1IJ9eEbbI=
=JFQA
-----END PGP SIGNATURE-----

--Apple-Mail=_CE91984A-912F-47A0-BC55-ED2CFEAABF0D--