Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:125736
Feedback-ID: ifab94697:Fastmail
Precedence: bulk
MIME-Version: 1.0
Date: Wed, 02 Oct 2024 22:02:13 +0200
To: =?UTF-8?Q?Alexandru_P=C4=83tr=C4=83nescu?= <drealecs@gmail.com>
Cc: "Juliette Reinders Folmer" <php-internals_nospam@adviesenzo.nl>,
 "PHP internals" <internals@lists.php.net>,
 "Ayesh Karunaratne" <ayesh@php.watch>
Message-ID: <31d49cf4-9e22-4e15-8589-ffc95a4ada82@app.fastmail.com>
In-Reply-To: 
 <CAAwdEzB+iHSmogoj=Yu4RVw8rY+B6Og_ORCMCg7LnoKxZ9cP6g@mail.gmail.com>
References: <66FC7B9F.5070906@adviesenzo.nl>
 <CAAwdEzC8sGt2+PcJfWTfLDooPTsqBk7YAdK-zXs7QMgDALX+LA@mail.gmail.com>
 <a2f73548-c6b7-4ad1-b124-176212050644@app.fastmail.com>
 <CAAwdEzB+iHSmogoj=Yu4RVw8rY+B6Og_ORCMCg7LnoKxZ9cP6g@mail.gmail.com>
Subject: Re: [PHP-DEV] [RFC] [Discussion] Add get_declared_enums() function
Content-Type: multipart/alternative;
 boundary=664b6b4cbba74bde8ac8d29fbcb299fa
From: rob@bottled.codes ("Rob Landers")

--664b6b4cbba74bde8ac8d29fbcb299fa
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

On Wed, Oct 2, 2024, at 14:27, Alexandru P=C4=83tr=C4=83nescu wrote:
>=20
> On Wed, Oct 2, 2024 at 11:06=E2=80=AFAM Rob Landers <rob@bottled.codes=
> wrote:
>> __
>> On Wed, Oct 2, 2024, at 09:20, Alexandru P=C4=83tr=C4=83nescu wrote:
>>>=20
>>> On Wed, Oct 2, 2024 at 1:48=E2=80=AFAM Juliette Reinders Folmer <php=
-internals_nospam@adviesenzo.nl> wrote:
>>>> L.S.,
>>>>=20
>>>> After the earlier discussion [1] regarding this topic in August, it=
 is our pleasure to present the RFC to add a `get_declared_enums()` func=
tion to PHP for discussion.
>>>>=20
>>>> https://wiki.php.net/rfc/get_declared_enums
>>>>=20
>>>> We look forward to your feedback and hope for a constructive discus=
sion.
>>>=20
>>> I'm all for `get_declared_enums()`.
>>> However, I don't accept that enums are not classes and going with th=
e assumption that they are not will bring more problems than it solves.
>>>=20
>>> I noticed how you analyzed things internally, that everything is a c=
lass internally, including an interface or a trait.
>>> But this is an API for the userland, and in userland this is not tru=
e.
>>> The 3 structures: classes, interfaces and traits and distinct things=
 that you can use in different ways in (generated) code.
>>>=20
>>> I believe that in userland, the common understanding is that enums a=
re just a specific type of class, final ones.
>>> I would find it surprising if `get_declared_classes()` would not ret=
urn enums.
>>>=20
>>> Avoiding a BC break would be nice also if there is no reason for it.
>>> One would be able to get only the enums by using something like:
>>> ```php
>>> function get_declared_enums_only() {
>>>     return array_diff(get_declared_classes(), get_declared_enums());
>>> }
>>> ```
>>>=20
>>> --=20
>>> Alex
>>>=20
>>=20
>> Interesting observation Alexandru,
>>=20
>> I find it weird that class_exists("MyEnum") would return true, but th=
ere is only one symbol table=E2=80=94you can=E2=80=99t name a class and =
enum with the same name, so it makes some sense. However, while enums ha=
ve =E2=80=9Csome=E2=80=9D similarities with classes in PHP, they are not=
 the same. For example, you cannot implement arbitrary interfaces (ie, S=
tringable), you cannot extend another class, you cannot instantiate them=
 with =E2=80=9Cnew=E2=80=9D, you cannot have state, you cannot clone the=
m, magic methods are forbidden, etc.
>>=20
>> You cannot build a =E2=80=9Cclass=E2=80=9D with these rules in PHP, s=
o I don=E2=80=99t think enums quack sufficiently like classes to be call=
ed a class. I would argue that it is an object, however. I think it woul=
d be worth implementing a get_declared_objects() that behaves like get_d=
eclared_classes() currently does.=20
>>=20
>> =E2=80=94 Rob
>=20
> Yes, that's a good point as well.
>=20
> I think my view comes from when I initially dug (15+ years ago) into w=
hat an enum in Java is, and learned that it is just a syntactic sugar, a=
nd a final class would be generated implementing `Comparable` and extend=
ing an abstract class `Enum`.
>=20
> And I think in PHP that could be a similar view. And most of the limit=
ations we have are not impossible to create with a standard class.
>=20
> I think in time we might get to remove some limitations.
> I, personally, don't agree with the `Stringable` related limitation.
> And also with the limitation on no state. In Java this is the simple w=
ay to create a singleton: an enum with one case; and we can't have this =
in PHP.
>=20
> If we remove the limitations, should we reclassify at that point enums=
 as classes?
>=20
> --=20
> Alex
> =20

Hey Alex,

IMHO, the only thing keeping an enum from being "sugared classes" is the=
 ability to define state. I'm not even sure why it is a limitation becau=
se it is relatively easy to "hack it in": https://3v4l.org/nqZfK=20

Being able to define state and behavior are fundamental parts of what ma=
kes a class a class. If you can't have state, then I'm not sure it is a =
class. We're getting into theoretical computer science here, but that is=
, I think, a generally accepted basic definition in Object Oriented Prog=
ramming.[1]

The rest could be explained as just sugar for:
 1. final class
 2. extend base enum class (hence why it can't extend another class)
 3. final clone method in base class which prevents cloning
 4. hidden magic method implementations that throw (hence unable to impl=
ement stringable)
 5. singletons for each case
 6. etc
(and before someone argues that it "isn't how it works", I am merely des=
cribing the sugar)

So, yes, once that restriction is lifted, I think we could officially ca=
ll them classes. Until then though, I think they're just "enums?"

On Wed, Oct 2, 2024, at 14:19, Derick Rethans wrote:
> On 2 October 2024 09:06:20 BST, Rob Landers <rob@bottled.codes> wrote:
>=20
> >You cannot build a =E2=80=9Cclass=E2=80=9D with these rules in PHP, s=
o I don=E2=80=99t think enums quack sufficiently like classes to be call=
ed a class. I would argue that it is an object, however. I think it woul=
d be worth implementing a get_declared_objects() that behaves like get_d=
eclared_classes() currently does.=20
>=20
> Objects are instantiated classes, and not declared elements, so that n=
ame wouldn't make sense.
>=20
> cheers
> Derick=20
>=20

This is also why I suggested calling them objects instead of classes. En=
um cases are instantiated enums which are objects, but they aren't class=
es. We could be more specific and call it get_declared_object_types()

I'd absolutely love to hear a reason, rooted in computer science, why PH=
P enums are classes. I'm no expert on the subject, but this has bothered=
 me for years... so I would love to learn more or discuss the topic furt=
her.

[1]: https://books.google.com/books?id=3D9NGWq3K1RwUC&pg=3DPA18 Foundati=
ons of Object Oriented Languages: types and semantics, Bruce, 2002 -- ju=
st had to break out an old text book :)

=E2=80=94 Rob
--664b6b4cbba74bde8ac8d29fbcb299fa
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}=0Ap.MsoNormal,p.MsoNoSpacing{margin:0}</=
style></head><body><div>On Wed, Oct 2, 2024, at 14:27, Alexandru P=C4=83=
tr=C4=83nescu wrote:<br></div><blockquote type=3D"cite" id=3D"qt" style=3D=
""><div dir=3D"ltr"><div dir=3D"ltr"><br></div><div class=3D"qt-gmail_qu=
ote"><div dir=3D"ltr" class=3D"qt-gmail_attr">On Wed, Oct 2, 2024 at 11:=
06=E2=80=AFAM Rob Landers &lt;rob@bottled.codes&gt; wrote:<br></div><blo=
ckquote class=3D"qt-gmail_quote" style=3D"margin-top:0px;margin-right:0p=
x;margin-bottom:0px;margin-left:0.8ex;border-left-width:1px;border-left-=
style:solid;border-left-color:rgb(204, 204, 204);padding-left:1ex;"><div=
 class=3D"qt-msg4813216649018816483"><div><u></u><br></div><div><div>On =
Wed, Oct 2, 2024, at 09:20, Alexandru P=C4=83tr=C4=83nescu wrote:<br></d=
iv><blockquote type=3D"cite" id=3D"qt-m_4813216649018816483qt"><div dir=3D=
"ltr"><div dir=3D"ltr"><br></div><div><div dir=3D"ltr">On Wed, Oct 2, 20=
24 at 1:48=E2=80=AFAM Juliette Reinders Folmer &lt;<a href=3D"mailto:php=
-internals_nospam@adviesenzo.nl" target=3D"_blank">php-internals_nospam@=
adviesenzo.nl</a>&gt; wrote:<br></div><blockquote style=3D"margin-top:0p=
x;margin-right:0px;margin-bottom:0px;margin-left:0.8ex;border-left-width=
:1px;border-left-style:solid;border-left-color:rgb(204, 204, 204);paddin=
g-left:1ex;"><div bgcolor=3D"#FFFFFF"><span><span class=3D"font" style=3D=
"font-family:Verdana;">L.S.,<br> <br> After the earlier discussion [1] r=
egarding this topic in August,=0A        it is our pleasure to present t=
he RFC to add a=0A        `get_declared_enums()` function to PHP for dis=
cussion.<br> <br> <a href=3D"https://wiki.php.net/rfc/get_declared_enums=
" target=3D"_blank">https://wiki.php.net/rfc/get_declared_enums</a><br> =
<br> We look forward to your feedback and hope for a constructive=0A    =
    discussion.</span></span></div></blockquote><div><br></div><div>I'm =
all for `get_declared_enums()`.<br></div><div>However, I don't accept th=
at enums are not classes and going with the assumption that they are not=
 will bring more problems than it solves.<br></div><div><br></div><div>I=
 noticed how&nbsp;you analyzed things internally, that everything is a c=
lass internally, including an interface or a trait.<br></div><div>But th=
is is an API for the userland, and in userland this is not true.<br></di=
v><div>The 3 structures: classes, interfaces and traits and distinct thi=
ngs that you can use in different ways in (generated) code.<br></div><di=
v><br></div><div>I believe that in userland, the common understanding is=
 that enums are just a specific type of class, final ones.<br></div><div=
>I would find it surprising if `get_declared_classes()` would not return=
 enums.<br></div><div><br></div><div>Avoiding a BC break would be nice a=
lso if there is no reason for it.<br></div><div>One would be able to get=
 only the enums by using something like:<br></div><div><div>```php<br></=
div><div>function get_declared_enums_only() {<br></div><div>&nbsp; &nbsp=
; return array_diff(get_declared_classes(), get_declared_enums());<br></=
div><div>}<br></div><div>```<br></div></div><div><br></div><div>--&nbsp;=
<br></div><div>Alex<br></div><div><br></div></div></div></blockquote><di=
v><br></div><div>Interesting observation&nbsp;Alexandru,<br></div><div><=
br></div><div>I find it weird that class_exists("MyEnum") would return t=
rue, but there is only one symbol table=E2=80=94you can=E2=80=99t name a=
 class and enum with the same name, so it makes some sense. However, whi=
le enums have =E2=80=9Csome=E2=80=9D similarities with classes in PHP, t=
hey are not the same. For example, you cannot implement arbitrary interf=
aces (ie, Stringable), you cannot extend another class, you cannot insta=
ntiate them with =E2=80=9Cnew=E2=80=9D, you cannot have state, you canno=
t clone them, magic methods are forbidden, etc.<br></div><div><br></div>=
<div>You cannot build a =E2=80=9Cclass=E2=80=9D with these rules in PHP,=
 so I don=E2=80=99t think enums quack sufficiently like classes to be ca=
lled a class. I would argue that it is an object, however. I think it wo=
uld be worth implementing a get_declared_objects() that behaves like get=
_declared_classes() currently does.&nbsp;<br></div><div><br></div><div i=
d=3D"qt-m_4813216649018816483sig121229152">=E2=80=94 Rob<br></div></div>=
</div></blockquote><div><br></div><div>Yes, that's a good point as well.=
<br></div><div><br></div><div>I think my view comes from when I initiall=
y dug (15+ years ago) into what an enum in Java is, and learned that it =
is just a syntactic sugar, and a final class would be generated implemen=
ting `Comparable` and extending an abstract class `Enum`.<br></div><div>=
<br></div><div>And I think in PHP that could be a similar view. And most=
 of the limitations we have are not impossible to create with a standard=
 class.<br></div><div><br></div><div>I think in time we might get to rem=
ove some limitations.<br></div><div>I, personally, don't agree with the =
`Stringable` related limitation.<br></div><div>And also with the limitat=
ion on no state. In Java this is the simple way to create a singleton: a=
n enum with one case; and we can't have this in PHP.<br></div><div><br><=
/div><div>If we remove the limitations, should we reclassify&nbsp;at tha=
t point enums as classes?<br></div><div><br></div><div>--&nbsp;<br></div=
><div>Alex<br></div><div>&nbsp;<br></div></div></div></blockquote><div><=
br></div><div>Hey Alex,<br></div><div><br></div><div>IMHO, the only thin=
g keeping an enum from being "sugared classes" is the ability to define =
state. I'm not even sure why it is a limitation because it is relatively=
 easy to "hack it in":&nbsp;<a href=3D"https://3v4l.org/nqZfK">https://3=
v4l.org/nqZfK</a>&nbsp;<br></div><div><br></div><div>Being able to defin=
e state and behavior are fundamental parts of what makes a class a class=
. If you can't have state, then I'm not sure it is a class. We're gettin=
g into theoretical computer science here, but that is, I think, a genera=
lly accepted basic definition in Object Oriented Programming.[1]<br></di=
v><div><br></div><div>The rest could be explained as just sugar for:<br>=
</div><ol><li>final class<br></li><li>extend base enum class (hence why =
it can't extend another class)<br></li><li>final clone method in base cl=
ass which prevents cloning<br></li><li>hidden magic method implementatio=
ns that throw (hence unable to implement stringable)<br></li><li>singlet=
ons for each case<br></li><li>etc<br></li></ol><div>(and before someone =
argues that it "isn't how it works", I am merely describing the sugar)<b=
r></div><div><br></div><div>So, yes, once that restriction is lifted, I =
think we could officially call them classes. Until then though, I think =
they're just "enums?"<br></div><div><br></div><div>On Wed, Oct 2, 2024, =
at 14:19, Derick Rethans wrote:<br></div><blockquote type=3D"cite" id=3D=
"qt" style=3D""><div>On 2 October 2024 09:06:20 BST, Rob Landers &lt;<a =
href=3D"mailto:rob@bottled.codes">rob@bottled.codes</a>&gt; wrote:<br></=
div><div><br></div><div>&gt;You cannot build a =E2=80=9Cclass=E2=80=9D w=
ith these rules in PHP, so I don=E2=80=99t think enums quack sufficientl=
y like classes to be called a class. I would argue that it is an object,=
 however. I think it would be worth implementing a get_declared_objects(=
) that behaves like get_declared_classes() currently does.&nbsp;<br></di=
v><div><br></div><div>Objects are instantiated classes, and not declared=
 elements, so that name wouldn't make sense.<br></div><div><br></div><di=
v>cheers<br></div><div>Derick&nbsp;<br></div><div><br></div></blockquote=
><div><br></div><div>This is also why I suggested calling them objects i=
nstead of classes. Enum cases are instantiated enums which are objects, =
but they aren't classes. We could be more specific and call it get_decla=
red_object_types()<br></div><div><br></div><div>I'd absolutely love to h=
ear a reason, rooted in computer science, why PHP enums are classes. I'm=
 no expert on the subject, but this has bothered me for years... so I wo=
uld love to learn more or discuss the topic further.</div><div><br></div=
><div>[1]:&nbsp;<a href=3D"https://books.google.com/books?id=3D9NGWq3K1R=
wUC&amp;pg=3DPA18">https://books.google.com/books?id=3D9NGWq3K1RwUC&amp;=
pg=3DPA18</a>&nbsp;Foundations of Object Oriented Languages: types and s=
emantics, Bruce, 2002 -- just had to break out an old text book :)<br></=
div><div><br></div><div id=3D"sig121229152">=E2=80=94 Rob<br></div></bod=
y></html>
--664b6b4cbba74bde8ac8d29fbcb299fa--