Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:126401
Precedence: bulk
MIME-Version: 1.0
References: <a995ac24905030e168e7f6006b996cd5@bastelstu.be>
 <CAMzRQTcTrZR2=vnjvVZRt=fEJpiBci9HN36BQAwG8+FK60Gc1A@mail.gmail.com>
 <5c142df3-94f8-45ba-b5c6-af3b4f7caa8b@varteg.nz> <CAJmjfvUQEEnEVpjFTr_Kkzsjq5f1bfwtPDkKPhg=hXq-0EpzsQ@mail.gmail.com>
 <fabda1c573ff3c605c2b8da3e74eabf5@bastelstu.be>
In-Reply-To: <fabda1c573ff3c605c2b8da3e74eabf5@bastelstu.be>
Date: Fri, 14 Feb 2025 10:10:04 +0100
Message-ID: <CAJmjfvXx5wvN5rHEHjEXdV+2U0RFL-tN7NsSQMVWjhhLPNo=3g@mail.gmail.com>
Subject: Re: [PHP-DEV] Re: RFC: Marking return values as important (#[\NoDiscard])
To: =?UTF-8?Q?Tim_D=C3=BCsterhus?= <tim@bastelstu.be>
Cc: volker@tideways-gmbh.com, internals@lists.php.net
Content-Type: multipart/alternative; boundary="000000000000289ee8062e168b3b"
From: jakob@givoni.dk (Jakob Givoni)

--000000000000289ee8062e168b3b
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Thu, Feb 13, 2025 at 3:17=E2=80=AFPM Tim D=C3=BCsterhus <tim@bastelstu.b=
e> wrote:

> Hi
>
> Am 2025-02-13 09:16, schrieb Jakob Givoni:
> > Attributes were added as a structured replacement for docblock props
> > and I
> > don't like it when they affect how a program actually runs (as long as
> > you're not using reflection).
>
> Excluding the `#[\Attribute]` attribute, PHP currently has 5 native
> attributes and they all affect how the program runs. The initial
> accepted Attribute RFC even lists several =E2=80=9Cbehavior-affecting=E2=
=80=9D
> attributes in the =E2=80=9CUse Cases=E2=80=9D section:
> https://wiki.php.net/rfc/attributes_v2#use_cases. It is probably fair to
> say that use-cases like `#[\NoDiscard]` do not go against the vision
> intended by the Attribute RFC.
>
> You could think of it as the PHP engine using Reflection internally to
> do something differently.
>

I think all the native attributes so far, with the exception of
#[\AllowDynamicProperties],
only affect the program at compile-time, or by emitting errors.
They don't affect the program otherwise by changing behavior,
throwing exceptions etc. (As long as you don't use reflection, nor promote
errors to exceptions.)

AllowDynamicProperties is the only one that may change the program behavior
at run-time by throwing an exception if the property is removed from a
class while running PHP versions 9.0+.
So this is not even a reality yet, since PHP 9 is not yet released.
And I think it sets a bad precedent. Perhaps AllowDynamicProperties should
have been a
language construct instead. Or only fail at compile time.

The original attributes RFC does mention "runtime behavior" in the
introduction,
but none of the use-cases proposed showcase it without using reflection or
promoting errors to exceptions (as far as I can see).

I don't mean to be rude and use your own words against you, but I found
this quote in the "Why an attribute and not a keyword?" section of
your #[\Override] RFC:

> This RFC proposes an attribute instead of a keyword, because contrary to
> other modifiers (e.g. visibility) that are part of the method signature,
> the attribute does not affect behavior or compatibility for users that
> further extend a given class and neither does it affect users that call t=
he
> method. It is purely an assistance to the author of a given class.

 (and I apologize if I misunderstand your intention or take it out of
context completely), but
I like the sentiment that "the attribute does not affect behavior or
compatibility for users"
and that it's part of the argument for an attribute instead of a keyword
(language construct).

I feel like #[NoDiscard] is getting dangerously close to being part of a
method signature,
especially if it throws exceptions at run-time (which I know it doesn't do
in your RFC, so I'm less against it).

I might be the only "purist" here, but just thought I would highlight that
at least one php developer
is hesitant with the direction attributes might be taking :-)

Best,
Jakob

--000000000000289ee8062e168b3b
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"ltr"><br></div><div class=3D"gmail_quote gmail=
_quote_container"><div dir=3D"ltr" class=3D"gmail_attr">On Thu, Feb 13, 202=
5 at 3:17=E2=80=AFPM Tim D=C3=BCsterhus &lt;<a href=3D"mailto:tim@bastelstu=
.be">tim@bastelstu.be</a>&gt; wrote:<br></div><blockquote class=3D"gmail_qu=
ote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,20=
4);padding-left:1ex">Hi<br>
<br>
Am 2025-02-13 09:16, schrieb Jakob Givoni:<br>
&gt; Attributes were added as a structured replacement for docblock props <=
br>
&gt; and I<br>
&gt; don&#39;t like it when they affect how a program actually runs (as lon=
g as<br>
&gt; you&#39;re not using reflection).<br>
<br>
Excluding the `#[\Attribute]` attribute, PHP currently has 5 native <br>
attributes and they all affect how the program runs. The initial <br>
accepted Attribute RFC even lists several =E2=80=9Cbehavior-affecting=E2=80=
=9D <br>
attributes in the =E2=80=9CUse Cases=E2=80=9D section: <br>
<a href=3D"https://wiki.php.net/rfc/attributes_v2#use_cases" rel=3D"norefer=
rer" target=3D"_blank">https://wiki.php.net/rfc/attributes_v2#use_cases</a>=
. It is probably fair to <br>
say that use-cases like `#[\NoDiscard]` do not go against the vision <br>
intended by the Attribute RFC.<br>
<br>
You could think of it as the PHP engine using Reflection internally to <br>
do something differently.<br></blockquote><div><br></div><div>I think all t=
he native attributes so far, with the exception of #[\AllowDynamicPropertie=
s],</div><div>only affect the program at compile-time, or by emitting error=
s.</div><div>They don&#39;t affect the program otherwise by changing behavi=
or,=C2=A0</div><div>throwing exceptions etc. (As long as you don&#39;t use =
reflection, nor promote errors to exceptions.)</div><div><br></div><div>All=
owDynamicProperties is the only one that may change the program behavior=C2=
=A0</div><div>at run-time by throwing an exception if the property is remov=
ed from a class while running PHP versions 9.0+.</div><div>So this is not e=
ven a reality yet, since PHP 9 is not yet released.</div><div>And I think i=
t sets a bad precedent. Perhaps AllowDynamicProperties should have been a=
=C2=A0</div><div>language construct instead. Or only fail at compile time.<=
/div><div><br></div><div>The original attributes RFC does mention &quot;run=
time behavior&quot; in the introduction,=C2=A0</div><div>but none of the us=
e-cases proposed showcase it without using reflection or=C2=A0</div><div>pr=
omoting errors to exceptions (as far as I can see).</div><div><br></div><di=
v>I don&#39;t mean to be rude and use your own words against you, but I fou=
nd this quote in the &quot;Why an attribute and not a keyword?&quot; sectio=
n of your=C2=A0#[\Override] RFC:<br><blockquote class=3D"gmail_quote" style=
=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding=
-left:1ex">This RFC proposes an attribute instead of a keyword, because con=
trary to other modifiers (e.g. visibility) that are part of the method sign=
ature, the attribute does not affect behavior or compatibility for users th=
at further extend a given class and neither does it affect users that call =
the method. It is purely an assistance to the author of a given class.</blo=
ckquote><div>=C2=A0(and I apologize if I misunderstand your intention or ta=
ke it out of context completely), but=C2=A0</div><div>I like the sentiment =
that &quot;the attribute does not affect behavior or compatibility for user=
s&quot;=C2=A0</div><div>and that it&#39;s part of the argument for an attri=
bute instead of a keyword (language construct).</div><div><br></div><div>I =
feel like #[NoDiscard] is getting dangerously close to being part of a meth=
od signature,=C2=A0</div><div>especially if it throws exceptions at run-tim=
e (which I know it doesn&#39;t do in your RFC, so I&#39;m less against it).=
</div><div><br></div><div><div>I might be the only &quot;purist&quot; here,=
 but just thought I would highlight that at least one php developer=C2=A0</=
div><div>is hesitant with the direction attributes might be taking :-)</div=
></div><div>=C2=A0</div></div><div>Best,</div><div>Jakob</div></div></div>

--000000000000289ee8062e168b3b--