Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:125931
Feedback-ID: ifab94697:Fastmail
Precedence: bulk
MIME-Version: 1.0
Date: Sat, 09 Nov 2024 10:35:01 +0100
To: internals@lists.php.net
Message-ID: <57517dcc-4f7f-4f3e-ac17-0d56a2ae760f@app.fastmail.com>
In-Reply-To: 
 <AS4P192MB162200EBAB671895CA643F3F985E2@AS4P192MB1622.EURP192.PROD.OUTLOOK.COM>
References: 
 <AS4P192MB162200EBAB671895CA643F3F985E2@AS4P192MB1622.EURP192.PROD.OUTLOOK.COM>
Subject: Re: [PHP-DEV] First-class types RFC
Content-Type: multipart/alternative;
 boundary=bd585698d73841e990401b755901e138
From: rob@bottled.codes ("Rob Landers")

--bd585698d73841e990401b755901e138
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

On Sat, Nov 9, 2024, at 09:39, Adam Kecskes wrote:
> Hello internals,
>=20
> I propose adding first-class types to PHP, as this feature could great=
ly enhance the language, especially given PHP's unique runtime type-chec=
king system. Many libraries currently rely on class strings as function =
parameters, which could be made significantly safer with native types ra=
ther than strings. This addition would not only streamline existing code=
 but also create a foundation for future language extensions. For instan=
ce, generic classes could be implemented as syntactic sugar, which would=
 translate into a class property holding the type parameter(s) for each =
instance.
>=20
> A minimal viable version could allow defining types as expressions, in=
troduce a new 'type' type, and expand is_a to accept those expressions l=
ike:
>=20
> is_a($object, \ClassOne | \ClassTwo);
>=20
> An example interface might look like:
>=20
> interface ContainerInterface
> {
>     public function get(type $type);
> }
>=20
> I=E2=80=99d appreciate your feedback on this idea before proceeding wi=
th an RFC or draft PR. Please let me know your thoughts!
>=20
> Best Regards,
> Adam Kecskes

Hey Adam,

While working on https://wiki.php.net/rfc/typed-aliases, (and an associa=
ted early discussion: https://externals.io/message/125455#125455), it wo=
uld make pattern matching almost trivial to implement, as well as first-=
class types. That being said, my life has been a bit of a whirlwind in t=
he last month, so I've had little time to work on these things. I'm hopi=
ng for things to settle down soon though. =F0=9F=AB=B0

I highly recommend seeking out the pattern matching RFC that is still be=
ing worked on as well: https://externals.io/message/123701#123701

Long story short, there is ongoing work towards what you are proposing h=
ere, but feel free to work on competing a competing RFC or contributing =
to ongoing discussions. Different ideas and approaches are almost always=
 good, IMHO.

=E2=80=94 Rob
--bd585698d73841e990401b755901e138
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE html><html><head><title></title><style type=3D"text/css">=0Ap.=
MsoNormal,p.MsoNoSpacing{margin:0}=0Ap.MsoNormal,p.MsoNoSpacing{margin:0=
}</style></head><body><div>On Sat, Nov 9, 2024, at 09:39, Adam Kecskes w=
rote:<br></div><blockquote type=3D"cite" id=3D"qt" style=3D""><div class=
=3D"qt-elementToProof" style=3D"font-family:Calibri, Helvetica, sans-ser=
if;font-size:12pt;color:rgb(0, 0, 0);">Hello internals,<br></div><div st=
yle=3D"font-family:Calibri, Helvetica, sans-serif;font-size:12pt;color:r=
gb(0, 0, 0);"><br></div><div style=3D"font-family:Calibri, Helvetica, sa=
ns-serif;font-size:12pt;color:rgb(0, 0, 0);">I propose adding first-clas=
s types to PHP, as this feature could greatly enhance the language, espe=
cially given PHP's unique runtime type-checking system. Many libraries c=
urrently rely on class strings as function parameters, which could be ma=
de significantly=0A safer with native types rather than strings. This ad=
dition would not only streamline existing code but also create a foundat=
ion for future language extensions. For instance, generic classes could =
be implemented as syntactic sugar, which would translate into=0A a class=
 property holding the type parameter(s) for each instance.<br></div><div=
 style=3D"font-family:Calibri, Helvetica, sans-serif;font-size:12pt;colo=
r:rgb(0, 0, 0);"><br></div><div style=3D"font-family:Calibri, Helvetica,=
 sans-serif;font-size:12pt;color:rgb(0, 0, 0);">A minimal viable version=
 could allow defining types as expressions, introduce a new 'type' type,=
 and expand is_a to accept those expressions like:<br></div><div style=3D=
"font-family:Calibri, Helvetica, sans-serif;font-size:12pt;color:rgb(0, =
0, 0);"><br></div><div style=3D"font-family:Calibri, Helvetica, sans-ser=
if;font-size:12pt;color:rgb(0, 0, 0);">is_a($object, \ClassOne | \ClassT=
wo);<br></div><div style=3D"font-family:Calibri, Helvetica, sans-serif;f=
ont-size:12pt;color:rgb(0, 0, 0);"><br></div><div style=3D"font-family:C=
alibri, Helvetica, sans-serif;font-size:12pt;color:rgb(0, 0, 0);">An exa=
mple interface might look like:<br></div><div style=3D"font-family:Calib=
ri, Helvetica, sans-serif;font-size:12pt;color:rgb(0, 0, 0);"><br></div>=
<div style=3D"font-family:Calibri, Helvetica, sans-serif;font-size:12pt;=
color:rgb(0, 0, 0);">interface ContainerInterface<br></div><div style=3D=
"font-family:Calibri, Helvetica, sans-serif;font-size:12pt;color:rgb(0, =
0, 0);">{<br></div><div style=3D"font-family:Calibri, Helvetica, sans-se=
rif;font-size:12pt;color:rgb(0, 0, 0);">&nbsp; &nbsp; public function ge=
t(type $type);<br></div><div style=3D"font-family:Calibri, Helvetica, sa=
ns-serif;font-size:12pt;color:rgb(0, 0, 0);">}<br></div><div style=3D"fo=
nt-family:Calibri, Helvetica, sans-serif;font-size:12pt;color:rgb(0, 0, =
0);"><br></div><div style=3D"font-family:Calibri, Helvetica, sans-serif;=
font-size:12pt;color:rgb(0, 0, 0);">I=E2=80=99d appreciate your feedback=
 on this idea before proceeding with an RFC or draft PR. Please let me k=
now your thoughts!<br></div><div style=3D"font-family:Calibri, Helvetica=
, sans-serif;font-size:12pt;color:rgb(0, 0, 0);"><br></div><div style=3D=
"font-family:Calibri, Helvetica, sans-serif;font-size:12pt;color:rgb(0, =
0, 0);">Best Regards,<br></div><div class=3D"qt-elementToProof" style=3D=
"font-family:Calibri, Helvetica, sans-serif;font-size:12pt;color:rgb(0, =
0, 0);">Adam Kecskes<br></div></blockquote><div><br></div><div>Hey Adam,=
<br></div><div><br></div><div>While working on&nbsp;<a href=3D"https://w=
iki.php.net/rfc/typed-aliases">https://wiki.php.net/rfc/typed-aliases</a=
>, (and an associated early discussion:&nbsp;<a href=3D"https://external=
s.io/message/125455#125455">https://externals.io/message/125455#125455</=
a>), it would make pattern matching almost trivial to implement, as well=
 as first-class types. That being said, my life has been a bit of a whir=
lwind in the last month, so I've had little time to work on these things=
. I'm hoping for things to settle down soon though. =F0=9F=AB=B0<br></di=
v><div><br></div><div>I highly recommend seeking out the pattern matchin=
g RFC that is still being worked on as well:&nbsp;<a href=3D"https://ext=
ernals.io/message/123701#123701">https://externals.io/message/123701#123=
701</a><br></div><div><br></div><div>Long story short, there is ongoing =
work towards what you are proposing here, but feel free to work on compe=
ting a competing RFC or contributing to ongoing discussions. Different i=
deas and approaches are almost always good, IMHO.</div><div><br></div><d=
iv id=3D"sig121229152">=E2=80=94 Rob<br></div></body></html>
--bd585698d73841e990401b755901e138--