Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:128637
Feedback-ID: ifab94697:Fastmail
Precedence: list
MIME-Version: 1.0
Date: Thu, 04 Sep 2025 16:50:08 +0200
To: internals@lists.php.net
Message-ID: <ee823501-40c5-455a-99b7-7a6c415ca1f4@app.fastmail.com>
In-Reply-To: <b7e870eb-9012-4fff-bd28-6a4a817a9181@app.fastmail.com>
References: <9c875e83-ddc8-4c5c-a368-747bef46e4a2@app.fastmail.com>
 <F60093E2-9310-4E6D-B1CA-C4B98D60B097@mabe.berlin>
 <25e79d21-7b7f-4d7f-b5cb-1e9cfd0aa657@varteg.nz>
 <b7e870eb-9012-4fff-bd28-6a4a817a9181@app.fastmail.com>
Subject: Re: [PHP-DEV] enum flag backed enum
Content-Type: multipart/alternative;
 boundary=efcfb98505114f1ea4a2023034f71917
From: rob@bottled.codes ("Rob Landers")

--efcfb98505114f1ea4a2023034f71917
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

On Thu, Sep 4, 2025, at 15:56, Larry Garfield wrote:
> On Wed, Sep 3, 2025, at 7:20 PM, Morgan wrote:
> > On 2025-09-04 09:23, Marc B. wrote:
> >>>
> >>> Attempting to use a backed value that is not a power-of-two would =
result in a compilation error. Bitwise operations are performed as thoug=
h they=E2=80=99re performed on a regular integer.
> >>>
> >>> What do you think?
> >>=20
> >> I think what we actually would need is an EnumSet which can be used=
 with any enum but (based on its ordinal) but this would play nicely tog=
ether with generics `new EnumSet<JsonOption>`.
> >>=20
> >
> > Agreed; using enums-backed-by-powers-of-two is just a way to encode =
sets=20
> > of values drawn from an enum (it's a way to enumerate the possible=20
> > subsets of a set). It should be possible to have a set of Enum value=
s=20
> > for any Enum, rather than break the Enum type concept by exposing=20
> > implementation details, and allow sets for some types but not others=20
> > depending on how they are serialised.
> >
> > A class can be built in userspace as a "SetOfEnum"; or such a class=20
> > could be added to the standard library; or even have, for each Enum=20
> > declared, an EnumSet (final) class is also implicitly declared.
>=20
> It feels like there's two different related topics here.
>=20
> One is using enums as a way to define a set of bit flags rather than c=
onstants.  I'm not sure if this is necessary, but if so, I'd expect a bi=
t more automation than the initial proposal here.  The statement that bi=
twise operations would fall back to integer ops makes me think it falls =
into the "fancy constants" category we have tried to avoid.
>=20
> The other is enum sets, which I am very much in favor of but designing=
 those in a clean way is non-trivial.  My collections research with Deri=
ck a while back included a `Set` type, which also supported operator ove=
rrides for the |, <, <=3D, -, and other operators, so it did effectively=
 give a very nice enum set syntax.  Whether that is the optimal approach=
 or not is unclear.  But an Enum Set feature should work with any enum. =
 (And of course runs into the eternal generics question.)
>=20
> So, Rob, which one are you really talking about?
>=20
> --Larry Garfield
>=20

Hey Larry, Marc, Morgan,

You are all asking basically the same question; I think: "This looks lik=
e a leaky abstraction and could possibly be solved more elegantly?"

I think this is a fair observation and a fair question; but I think it i=
s important not to have "magic". The power-of-two rule is to make it pos=
sible to work back how $enum->value =3D=3D=3D 15 (0x1111) even if you ar=
e completely new to the language. If you just use some magical cardinal =
order, it is impossible to reserve ranges, handle communications with ex=
ternal systems, etc. I'm not opposed to it, however, especially when end=
-to-end values are in PHP itself. But, when you need to communicate with=
 external systems, being able to have that control is *very* important.

After doing a bit more digging, it appears this can be handled entirely =
as an extension, so I think I'll just go that route for now.

=E2=80=94 Rob
--efcfb98505114f1ea4a2023034f71917
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE html><html><head><title></title></head><body><div>On Thu, Sep =
4, 2025, at 15:56, Larry Garfield wrote:</div><blockquote type=3D"cite" =
id=3D"qt" style=3D""><div>On Wed, Sep 3, 2025, at 7:20 PM, Morgan wrote:=
</div><div>&gt; On 2025-09-04 09:23, Marc B. wrote:</div><div>&gt;&gt;&g=
t;</div><div>&gt;&gt;&gt; Attempting to use a backed value that is not a=
 power-of-two would result in a compilation error. Bitwise operations ar=
e performed as though they=E2=80=99re performed on a regular integer.</d=
iv><div>&gt;&gt;&gt;</div><div>&gt;&gt;&gt; What do you think?</div><div=
>&gt;&gt;&nbsp;</div><div>&gt;&gt; I think what we actually would need i=
s an EnumSet which can be used with any enum but (based on its ordinal) =
but this would play nicely together with generics `new EnumSet&lt;JsonOp=
tion&gt;`.</div><div>&gt;&gt;&nbsp;</div><div>&gt;</div><div>&gt; Agreed=
; using enums-backed-by-powers-of-two is just a way to encode sets&nbsp;=
</div><div>&gt; of values drawn from an enum (it's a way to enumerate th=
e possible&nbsp;</div><div>&gt; subsets of a set). It should be possible=
 to have a set of Enum values&nbsp;</div><div>&gt; for any Enum, rather =
than break the Enum type concept by exposing&nbsp;</div><div>&gt; implem=
entation details, and allow sets for some types but not others&nbsp;</di=
v><div>&gt; depending on how they are serialised.</div><div>&gt;</div><d=
iv>&gt; A class can be built in userspace as a "SetOfEnum"; or such a cl=
ass&nbsp;</div><div>&gt; could be added to the standard library; or even=
 have, for each Enum&nbsp;</div><div>&gt; declared, an EnumSet (final) c=
lass is also implicitly declared.</div><div><br></div><div>It feels like=
 there's two different related topics here.</div><div><br></div><div>One=
 is using enums as a way to define a set of bit flags rather than consta=
nts.&nbsp; I'm not sure if this is necessary, but if so, I'd expect a bi=
t more automation than the initial proposal here.&nbsp; The statement th=
at bitwise operations would fall back to integer ops makes me think it f=
alls into the "fancy constants" category we have tried to avoid.</div><d=
iv><br></div><div>The other is enum sets, which I am very much in favor =
of but designing those in a clean way is non-trivial.&nbsp; My collectio=
ns research with Derick a while back included a `Set` type, which also s=
upported operator overrides for the |, &lt;, &lt;=3D, -, and other opera=
tors, so it did effectively give a very nice enum set syntax.&nbsp; Whet=
her that is the optimal approach or not is unclear.&nbsp; But an Enum Se=
t feature should work with any enum.&nbsp; (And of course runs into the =
eternal generics question.)</div><div><br></div><div>So, Rob, which one =
are you really talking about?</div><div><br></div><div>--Larry Garfield<=
/div><div><br></div></blockquote><div><br></div><div>Hey Larry, Marc, Mo=
rgan,</div><div><br></div><div>You are all asking basically the same que=
stion; I think: "This looks like a leaky abstraction and could possibly =
be solved more elegantly?"</div><div><br></div><div>I think this is a fa=
ir observation and a fair question; but I think it is important not to h=
ave "magic". The power-of-two rule is to make it possible to work back h=
ow $enum-&gt;value =3D=3D=3D 15 (0x1111) even if you are completely new =
to the language. If you just use some magical cardinal order, it is impo=
ssible to reserve ranges, handle communications with external systems, e=
tc. I'm not opposed to it, however, especially when end-to-end values ar=
e in PHP itself. But, when you need to communicate with external systems=
, being able to have that control is&nbsp;<i>very</i> important.<br></di=
v><div><br></div><div>After doing a bit more digging, it appears this ca=
n be handled entirely as an extension, so I think I'll just go that rout=
e for now.</div><div><br></div><div id=3D"sig121229152">=E2=80=94 Rob</d=
iv></body></html>
--efcfb98505114f1ea4a2023034f71917--