Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:122512
Feedback-ID: id5114917:Fastmail
Date: Tue, 27 Feb 2024 09:49:35 +0000
To: internals@lists.php.net, Frederik Bosch <f.bosch@genkgo.nl>
Subject: Re: [PHP-DEV] [RFC[ Property accessor hooks, take 2
User-Agent: K-9 Mail for Android
In-Reply-To: <963f5cc5-cdb1-4384-b519-5cb15640654e@genkgo.nl>
References: <fca97d6c-d4ec-4000-958a-72c32beb8421@app.fastmail.com> <CAJp_myUayeQ=irpFUG0OrUj=PjE3_mx395h_ARSdLmdyEB6_EQ@mail.gmail.com> <790b5b4e-f51b-4050-a12a-5fa903d0568f@app.fastmail.com> <52C6F501-8E23-42D7-8541-88A22AD79375@koalephant.com> <36e90d8d-d275-4ce9-9dd9-1e2422c6d3a9@app.fastmail.com> <2fdf1933-b51c-40cc-8d02-31899b96c71c@genkgo.nl> <d4832238-7af4-4bd4-b8a1-8ab8d56d51d2@rwec.co.uk> <95e93cb9-3ab0-4cf3-8ec5-83e74c9dd607@genkgo.nl> <876aff9f-3eae-4d2d-8e3f-30dfbbeed49c@rwec.co.uk> <963f5cc5-cdb1-4384-b519-5cb15640654e@genkgo.nl>
Message-ID: <58B82A81-8A89-4F17-B982-7FC36404032E@rwec.co.uk>
Precedence: bulk
MIME-Version: 1.0
Content-Type: multipart/alternative;
 boundary=----THPRZVTWNDWO5NK9O5RCSVN5SO3GES
Content-Transfer-Encoding: 7bit
From: imsop.php@rwec.co.uk ("Rowan Tommins [IMSoP]")

------THPRZVTWNDWO5NK9O5RCSVN5SO3GES
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable

On 26 February 2024 23:11:16 GMT, Frederik Bosch <f=2Ebosch@genkgo=2Enl> wr=
ote:

>And what happens in the following situation, how are multiple get calls w=
orking together?
>
>public string $fullName {
>=C2=A0=C2=A0=C2=A0 get =3D> $this->first =2E ' ' =2E $this->last; // is t=
his accessing the backed value, or is it accessing via get
>=C2=A0=C2=A0=C2=A0 set($value) =3D> $this->fullName =3D $value;
>}
>
>public string $first {
>=C2=A0=C2=A0=C2=A0 get =3D> explode(' ', $this->fullName)[0], // is this =
accessing the backed value, or is it accessing via get
>=C2=A0=C2=A0=C2=A0 set($value) =3D> $value;
>}

I don't think it's *that* confusing - the rule is not "hooks vs methods", =
it's "special access inside the property's own hook"=2E But as I say, I'm c=
oming around to the idea that using a different name for that "backing fiel=
d" / "raw value" might be sensible=2E


>> What would happen if a setter contained both "return 42;" and "return;"=
? The latter is explicitly allowed in "void" functions, but is also allowed=
 in a non-void function as meaning "return null;"
>return 42; // returns (int)42
>return; // early return, void, same as no return
>return null; // returns null

I'm not sure if you misunderstood my question, or just the context of why =
I asked it=2E I'm talking about a hook like this:

set($value) { if ($value) { return 42; } else { return; } }

Currently, the only definition of "void" in the language is that a void fu=
nction must not contain an explicit return value=2E We could turn that chec=
k around, and deduce that a certain hook is void=2E This hook would not pas=
s that check, so we would compile it to have an assignment, and the false c=
ase would assign null to the property=2E To avoid that, we would need some =
additional analysis to prove that in all possible paths, a return statement=
 with a value is reached=2E

The alternative would be to run the code, and somehow observe that it "ret=
urned void"=2E But "void" isn't a value we can represent at run-time; we wo=
uld need to set the return value to some special value just for this specif=
ic case=2E We would have to turn that on just for hook bodies, as returning=
 it from normal functions would be a huge BC break, and also not very usefu=
l - with union types, there would be plenty of better options for a functio=
n to indicate a return value that needs special handling=2E



>$generator =3D setCall($class, 'first', $value);
>foreach ($generator as $value) {
>=C2=A0=C2=A0 writeProperty($class, 'first', $value);
>}
>if ($generator->hasReturn()) {
>writeProperty($class, 'first', $generator->getReturn());
>}


That's already an order of magnitude more complicated than "the return val=
ue is used on the right-hand side of an assignment", and it's missing at le=
ast one case: set($value) { return $value; } will not compile to a generato=
r, so needs to skip and assign the value directly=2E

By "magic", what I meant was "hidden logic underneath that makes it work"=
=2E Assign-by-return has a small amount of magic - you can express it in ha=
lf a line of code; assign-by-yield has much more magic - a whole bunch of l=
oops and conditionals to operate your coroutine=2E



> The yield is much more intuitive than magic fields

I think we'll just have to differ in opinion on that one=2E Maybe you're j=
ust more used to working with coroutines than I am=2E

Note that yield also doesn't solve how to read the current backing value i=
n a get hook (or a set hook that wants to compare before and after), so we =
still need some way to refer to it=2E

Regards,
Rowan Tommins
[IMSoP]
------THPRZVTWNDWO5NK9O5RCSVN5SO3GES
Content-Type: text/html;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE html><html><body><div dir=3D"auto">On 26 February 2024 23:11:16 G=
MT, Frederik Bosch &lt;f=2Ebosch@genkgo=2Enl&gt; wrote:<br><br>&gt;And what=
 happens in the following situation, how are multiple get calls working tog=
ether?<br>&gt;<br>&gt;public string $fullName {<br>&gt;=C2=A0=C2=A0=C2=A0 g=
et =3D&gt; $this-&gt;first =2E ' ' =2E $this-&gt;last; // is this accessing=
 the backed value, or is it accessing via get<br>&gt;=C2=A0=C2=A0=C2=A0 set=
($value) =3D&gt; $this-&gt;fullName =3D $value;<br>&gt;}<br>&gt;<br>&gt;pub=
lic string $first {<br>&gt;=C2=A0=C2=A0=C2=A0 get =3D&gt; explode(' ', $thi=
s-&gt;fullName)[0], // is this accessing the backed value, or is it accessi=
ng via get<br>&gt;=C2=A0=C2=A0=C2=A0 set($value) =3D&gt; $value;<br>&gt;}<b=
r><br>I don't think it's *that* confusing - the rule is not "hooks vs metho=
ds", it's "special access inside the property's own hook"=2E But as I say, =
I'm coming around to the idea that using a different name for that "backing=
 field" / "raw value" might be sensible=2E<br><br><br>&gt;&gt; What would h=
appen if a setter contained both "return 42;" and "return;"? The latter is =
explicitly allowed in "void" functions, but is also allowed in a non-void f=
unction as meaning "return null;"<br>&gt;return 42; // returns (int)42<br>&=
gt;return; // early return, void, same as no return<br>&gt;return null; // =
returns null<br><br>I'm not sure if you misunderstood my question, or just =
the context of why I asked it=2E I'm talking about a hook like this:<br><br=
>set($value) { if ($value) { return 42; } else { return; } }<br><br>Current=
ly, the only definition of "void" in the language is that a void function m=
ust not contain an explicit return value=2E We could turn that check around=
, and deduce that a certain hook is void=2E This hook would not pass that c=
heck, so we would compile it to have an assignment, and the false case woul=
d assign null to the property=2E To avoid that, we would need some addition=
al analysis to prove that in all possible paths, a return statement with a =
value is reached=2E<br><br>The alternative would be to run the code, and so=
mehow observe that it "returned void"=2E But "void" isn't a value we can re=
present at run-time; we would need to set the return value to some special =
value just for this specific case=2E We would have to turn that on just for=
 hook bodies, as returning it from normal functions would be a huge BC brea=
k, and also not very useful - with union types, there would be plenty of be=
tter options for a function to indicate a return value that needs special h=
andling=2E<br><br><br><br>&gt;$generator =3D setCall($class, 'first', $valu=
e);<br>&gt;foreach ($generator as $value) {<br>&gt;=C2=A0=C2=A0 writeProper=
ty($class, 'first', $value);<br>&gt;}<br>&gt;if ($generator-&gt;hasReturn()=
) {<br>&gt;writeProperty($class, 'first', $generator-&gt;getReturn());<br>&=
gt;}<br><br><br>That's already an order of magnitude more complicated than =
"the return value is used on the right-hand side of an assignment", and it'=
s missing at least one case: set($value) { return $value; } will not compil=
e to a generator, so needs to skip and assign the value directly=2E<br><br>=
By "magic", what I meant was "hidden logic underneath that makes it work"=
=2E Assign-by-return has a small amount of magic - you can express it in ha=
lf a line of code; assign-by-yield has much more magic - a whole bunch of l=
oops and conditionals to operate your coroutine=2E<br><br><br><br>&gt; The =
yield is much more intuitive than magic fields<br><br>I think we'll just ha=
ve to differ in opinion on that one=2E Maybe you're just more used to worki=
ng with coroutines than I am=2E<br><br>Note that yield also doesn't solve h=
ow to read the current backing value in a get hook (or a set hook that want=
s to compare before and after), so we still need some way to refer to it=2E=
<br><br>Regards,</div><div dir=3D"auto">Rowan Tommins<br>[IMSoP]</div></bod=
y></html>
------THPRZVTWNDWO5NK9O5RCSVN5SO3GES--