Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:124976
Feedback-ID: ifab94697:Fastmail
Precedence: bulk
MIME-Version: 1.0
Date: Fri, 16 Aug 2024 16:27:46 +0200
To: internals@lists.php.net
Message-ID: <023ae126-af2e-4bc4-966d-636040d75983@app.fastmail.com>
In-Reply-To: <79817594-f9a9-4f02-b48a-4bf39e4372e6@app.fastmail.com>
References: <34548342633f1d87b2fc77c643bd895a18f71b66.camel@ageofdream.com>
 <CAAwdEzBWN40q10Q5ua_ZME29dJSqs=BHB1LxJ62mN-vXWNw0xQ@mail.gmail.com>
 <79817594-f9a9-4f02-b48a-4bf39e4372e6@app.fastmail.com>
Subject: Re: [PHP-DEV] [DISCUSSION] Class Constant Enums?
Content-Type: multipart/alternative;
 boundary=ecc9128cad224c1a900a61865f3cdd69
From: rob@bottled.codes ("Rob Landers")

--ecc9128cad224c1a900a61865f3cdd69
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

On Fri, Aug 16, 2024, at 16:17, Larry Garfield wrote:
> On Fri, Aug 16, 2024, at 6:35 AM, Alexandru P=C4=83tr=C4=83nescu wrote:
> > Hi Nick,
> >>=20
> >> Is there any interest in having enums as class constants?
> >>=20
> >> I'm often finding cases where I would like to have an enum inside o=
f a
> >> class, but don't want a free-floating enum that's basically like
> >> another class.
> >>=20
> >=20
> > ...<snip>...=20
> >=20
> >>=20
> >> class SSHClient {
> >>=20
> >>    public const enum CommandResult
> >>    {
> >>        case Success;
> >>        case Failure;
> >>        case Unknown;
> >>        case Timeout;
> >>    }
> >>=20
> >>    // ...
> >> }
> >>=20
> >>=20
> >> // Usage:
> >>=20
> >> SSHClient::CommandResult::Success
> >
> >
> > I feel this topic could be maybe more broad and be called "nested=20
> > classes" that are already supported in multiple languages: Java, Swi=
ft,=20
> > Python, C#, C++, JavaScript, etc.
> >
> > The syntax you showed is usually identical with what other languages=20
> > use, except that probably the const is unnecessary.
> > The nested class can have visibility as sometimes having it private=20
> > makes sense.
> > Accessing it through `::` is probably fine, but a deeper look at the=20
> > grammar might be necessary.
> > The nested class would have access to parent class private propertie=
s=20
> > and methods.
> >
> > I also mentioned this topic on the subject of defining a type in an=20
> > autoloader compatible way.
> > And indeed, a type could also be defined nested in a class if we wan=
t=20
> > to support that as well.
> >
> > Now, this feature is not simple, and I think it needs proper=20
> > sponsorship from someone experienced with internals.
> >
> > Regards,
> > Alex
>=20
> I agree with Alexandru.  Since enums are 90% syntactic sugar over clas=
ses, "inner enums" would be 80% of the way to "inner classes".  And I wo=
uld be in favor of inner classes. :-)  There's a lot of potential benefi=
ts there, but also a lot of edge cases to sort out regarding visibility,=
 what is allowed to extend from what, etc.  But that would support inner=
 enums as well.

From recently looking into this for totally unrelated reasons, nested en=
ums would be far easier to implement on a grammar level. Enums also have=
 some constraints that make it simpler than the general =E2=80=9Cnested =
classes,=E2=80=9D such as rules regarding inheritance.

As for the actual implementation, it=E2=80=99ll be the edges that kill y=
ou.

I would recommend just doing enums, and keep the scope smaller.=20

--ecc9128cad224c1a900a61865f3cdd69
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE html><html><head><title></title><style type=3D"text/css">p.Mso=
Normal,p.MsoNoSpacing{margin:0}</style></head><body><div>On Fri, Aug 16,=
 2024, at 16:17, Larry Garfield wrote:<br></div><blockquote type=3D"cite=
" id=3D"qt" style=3D""><div>On Fri, Aug 16, 2024, at 6:35 AM, Alexandru =
P=C4=83tr=C4=83nescu wrote:<br></div><div>&gt; Hi Nick,<br></div><div>&g=
t;&gt;&nbsp;<br></div><div>&gt;&gt; Is there any interest in having enum=
s as class constants?<br></div><div>&gt;&gt;&nbsp;<br></div><div>&gt;&gt=
; I'm often finding cases where I would like to have an enum inside of a=
<br></div><div>&gt;&gt; class, but don't want a free-floating enum that'=
s basically like<br></div><div>&gt;&gt; another class.<br></div><div>&gt=
;&gt;&nbsp;<br></div><div>&gt;&nbsp;<br></div><div>&gt; ...&lt;snip&gt;.=
..&nbsp;<br></div><div>&gt;&nbsp;<br></div><div>&gt;&gt;&nbsp;<br></div>=
<div>&gt;&gt; class SSHClient {<br></div><div>&gt;&gt;&nbsp;<br></div><d=
iv>&gt;&gt;&nbsp;&nbsp;&nbsp; public const enum CommandResult<br></div><=
div>&gt;&gt;&nbsp;&nbsp;&nbsp; {<br></div><div>&gt;&gt;&nbsp;&nbsp;&nbsp=
;&nbsp;&nbsp;&nbsp;&nbsp; case Success;<br></div><div>&gt;&gt;&nbsp;&nbs=
p;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; case Failure;<br></div><div>&gt;&gt;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; case Unknown;<br></div><div>&gt;=
&gt;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; case Timeout;<br></div><d=
iv>&gt;&gt;&nbsp;&nbsp;&nbsp; }<br></div><div>&gt;&gt;&nbsp;<br></div><d=
iv>&gt;&gt;&nbsp;&nbsp;&nbsp; // ...<br></div><div>&gt;&gt; }<br></div><=
div>&gt;&gt;&nbsp;<br></div><div>&gt;&gt;&nbsp;<br></div><div>&gt;&gt; /=
/ Usage:<br></div><div>&gt;&gt;&nbsp;<br></div><div>&gt;&gt; SSHClient::=
CommandResult::Success<br></div><div>&gt;<br></div><div>&gt;<br></div><d=
iv>&gt; I feel this topic could be maybe more broad and be called "neste=
d&nbsp;<br></div><div>&gt; classes" that are already supported in multip=
le languages: Java, Swift,&nbsp;<br></div><div>&gt; Python, C#, C++, Jav=
aScript, etc.<br></div><div>&gt;<br></div><div>&gt; The syntax you showe=
d is usually identical with what other languages&nbsp;<br></div><div>&gt=
; use, except that probably the const is unnecessary.<br></div><div>&gt;=
 The nested class can have visibility as sometimes having it private&nbs=
p;<br></div><div>&gt; makes sense.<br></div><div>&gt; Accessing it throu=
gh `::` is probably fine, but a deeper look at the&nbsp;<br></div><div>&=
gt; grammar might be necessary.<br></div><div>&gt; The nested class woul=
d have access to parent class private properties&nbsp;<br></div><div>&gt=
; and methods.<br></div><div>&gt;<br></div><div>&gt; I also mentioned th=
is topic on the subject of defining a type in an&nbsp;<br></div><div>&gt=
; autoloader compatible way.<br></div><div>&gt; And indeed, a type could=
 also be defined nested in a class if we want&nbsp;<br></div><div>&gt; t=
o support that as well.<br></div><div>&gt;<br></div><div>&gt; Now, this =
feature is not simple, and I think it needs proper&nbsp;<br></div><div>&=
gt; sponsorship from someone experienced with internals.<br></div><div>&=
gt;<br></div><div>&gt; Regards,<br></div><div>&gt; Alex<br></div><div><b=
r></div><div>I agree with Alexandru.&nbsp; Since enums are 90% syntactic=
 sugar over classes, "inner enums" would be 80% of the way to "inner cla=
sses".&nbsp; And I would be in favor of inner classes. :-)&nbsp; There's=
 a lot of potential benefits there, but also a lot of edge cases to sort=
 out regarding visibility, what is allowed to extend from what, etc.&nbs=
p; But that would support inner enums as well.<br></div></blockquote><di=
v><br></div><div>From recently looking into this for totally unrelated r=
easons, nested enums would be far easier to implement on a grammar level=
. Enums also have some constraints that make it simpler than the general=
 =E2=80=9Cnested classes,=E2=80=9D such as rules regarding inheritance.<=
br></div><div><br></div><div>As for the actual implementation, it=E2=80=99=
ll be the edges that kill you.<br></div><div><br></div><div>I would reco=
mmend just doing enums, and keep the scope smaller.&nbsp;</div><div><br>=
</div></body></html>
--ecc9128cad224c1a900a61865f3cdd69--