Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:128367
Feedback-ID: ifab94697:Fastmail
Precedence: list
MIME-Version: 1.0
Date: Sat, 02 Aug 2025 11:07:21 +0200
To: internals@lists.php.net
Message-ID: <2d516e15-2fc7-4ac4-b9c2-7146ac01cfb1@app.fastmail.com>
In-Reply-To: <c24d84f3-0d9a-4412-8ea5-a87ebac2e64a@jnvsor.net>
References: <c24d84f3-0d9a-4412-8ea5-a87ebac2e64a@jnvsor.net>
Subject: Re: [PHP-DEV] Protected inheritance hierarchies
Content-Type: multipart/alternative;
 boundary=9a7852dfe5a34314be65ab46820f93f7
From: rob@bottled.codes ("Rob Landers")

--9a7852dfe5a34314be65ab46820f93f7
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

On Tue, Jul 29, 2025, at 20:11, Jonathan Vollebregt wrote:
> I came across this edge case today:
>=20
> https://3v4l.org/R3Q8D
>=20
> Both psalm and phpstan think this code is A-OK (Once you add the=20
> requisite type hints) but it causes fatal errors way back to PHP 5.0.0
>=20
> I believe classes should be able to access protected properties on the=
ir=20
> siblings if the property in question was declared by a shared parent,=20
> but it seems both the engine and reflection (ie. getDeclaringClass)=20
> think that redeclaring a protected property makes it a property of the=20
> child, not the parent.
>=20
> This is particularly confusing since the parent class *can* access the=20
> child class' redeclared protected property, only the sibling can't.
>=20
> Properties were invariant until the introduction of property hooks, so=20
> the only edge cases I can think of would be in property hooks (But whe=
n=20
> the input is correctly typed this shouldn't be a problem either)
>=20
> Is there a technical reason for this behavior or would it be possible =
to=20
> relax this?
>=20
> - Jonathan
>=20

It's not an edge case, in C2, you redefine a protected variable with the=
 same name and shadowed the original $v. That $v is different than C's $=
v. It's easiest to see this with static access: https://3v4l.org/0SRWb#v=
8.4.10

However, I don't know of any way to unshadow a property from $this to ac=
cess the ancestor's value (other than using private access), but it exis=
ts and takes up memory; just accessing it is the hard part.

=E2=80=94 Rob
--9a7852dfe5a34314be65ab46820f93f7
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE html><html><head><title></title></head><body><div>On Tue, Jul =
29, 2025, at 20:11, Jonathan Vollebregt wrote:</div><blockquote type=3D"=
cite" id=3D"qt" style=3D""><div>I came across this edge case today:</div=
><div><br></div><div><a href=3D"https://3v4l.org/R3Q8D">https://3v4l.org=
/R3Q8D</a></div><div><br></div><div>Both psalm and phpstan think this co=
de is A-OK (Once you add the&nbsp;</div><div>requisite type hints) but i=
t causes fatal errors way back to PHP 5.0.0</div><div><br></div><div>I b=
elieve classes should be able to access protected properties on their&nb=
sp;</div><div>siblings if the property in question was declared by a sha=
red parent,&nbsp;</div><div>but it seems both the engine and reflection =
(ie. getDeclaringClass)&nbsp;</div><div>think that redeclaring a protect=
ed property makes it a property of the&nbsp;</div><div>child, not the pa=
rent.</div><div><br></div><div>This is particularly confusing since the =
parent class *can* access the&nbsp;</div><div>child class' redeclared pr=
otected property, only the sibling can't.</div><div><br></div><div>Prope=
rties were invariant until the introduction of property hooks, so&nbsp;<=
/div><div>the only edge cases I can think of would be in property hooks =
(But when&nbsp;</div><div>the input is correctly typed this shouldn't be=
 a problem either)</div><div><br></div><div>Is there a technical reason =
for this behavior or would it be possible to&nbsp;</div><div>relax this?=
</div><div><br></div><div>- Jonathan</div><div><br></div></blockquote><d=
iv><br></div><div>It's not an edge case, in C2, you redefine a protected=
 variable with the same name and shadowed the original $v. That $v is di=
fferent than C's $v. It's easiest to see this with static access:&nbsp;<=
a href=3D"https://3v4l.org/0SRWb#v8.4.10">https://3v4l.org/0SRWb#v8.4.10=
</a><br></div><div><br></div><div>However, I don't know of any way to un=
shadow a property from $this to access the ancestor's value (other than =
using private access), but it exists and takes up memory; just accessing=
 it is the hard part.</div><div><br></div><div id=3D"sig121229152">=E2=80=
=94 Rob</div></body></html>
--9a7852dfe5a34314be65ab46820f93f7--