Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:125083
Precedence: bulk
MIME-Version: 1.0
References: <1b59392a-68cb-36eb-0fef-977ac7113520@php.net> <AM8P250MB01706B1541E1283F4065B94AE28C2@AM8P250MB0170.EURP250.PROD.OUTLOOK.COM>
 <CAP1Jc11djBvAkWgMwM-SzUQyW4AQ+TUDtLNKmyFNCmbJvKW+WA@mail.gmail.com>
In-Reply-To: <CAP1Jc11djBvAkWgMwM-SzUQyW4AQ+TUDtLNKmyFNCmbJvKW+WA@mail.gmail.com>
Date: Tue, 20 Aug 2024 22:09:55 -0700
Message-ID: <CAMrTa2HF3CM0gFom-M_E=j62kBxo_m+o7w2_a2UPYnKKEGLg8A@mail.gmail.com>
Subject: Re: [PHP-DEV] State of Generics and Collections
To: Arnaud Le Blanc <arnaud.lb@gmail.com>
Cc: Bob Weinand <bobwei9@hotmail.com>, Derick Rethans <derick@php.net>, 
	PHP Developers Mailing List <internals@lists.php.net>
Content-Type: multipart/alternative; boundary="000000000000475b1206202a8ec2"
From: jordan.ledoux@gmail.com (Jordan LeDoux)

--000000000000475b1206202a8ec2
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Tue, Aug 20, 2024 at 6:02=E2=80=AFAM Arnaud Le Blanc <arnaud.lb@gmail.co=
m> wrote:

> Hi Bob,
>
> On Tue, Aug 20, 2024 at 12:18=E2=80=AFAM Bob Weinand <bobwei9@hotmail.com=
> wrote:
> > The fluid Arrays section says "A PoC has been implemented, but the
> performance impact is still uncertain". Where may I find that PoC for my
> curiosity? I'm imagining the implementation of the array types as a count=
ed
> collection of types of the entries. But without the PoC I may only guess.
>
> I may publish the PoC at some point, but in the meantime here is a
> short description of how it's implemented:
>
> - The zend_array has a zend_type member representing the type of its
> elements
> - Everytime we add or update a member, we union its type with the
> array type. For simple types it's just a |=3D operation. For arrays with
> a single class it's also simple. For complex types it's more expensive
> currently, but it may be possible to cache transitions to make this
> cheaper.
> - Updating the array type on deletes requires to either maintain a
> counter of every type, or to re-compute the type entirely everytime.
> Both are probably too expensive. Instead, we don't update the type on
> deletes, but we re-compute the type entirely when a type check fails.
> This is based on two hypotheses: 1. A delete rarely changes an array's
> type in practice, and 2. Type checks rarely fail
> - References are treated as mixed, so adding a reference to an array
> or taking a reference to an element changes its type to mixed. Passing
> an array<mixed> to a more specific array<something> will cause a
> re-compute, which also de-refs every reference.
> - Updating a nested element requires updating the type of every parent
>
> > It also says "Another issue is that [...] typed properties may not be
> possible.". Why would that be the case? Essentially a typed property woul=
d
> just be a static array, which you describe in the section right below.
>
> It becomes complicated when arrays contain references or nested
> arrays. Type constraints must be propagated to nested arrays, but also
> removed when an array is not reachable via a typed property anymore.
>
> E.g.
>
> class C {
>     public array<array<int>> $prop;
> }
>
> $a =3D &$c->prop[0];
> $a[] =3D 'string'; // must be an error
> unset($c->prop[0]);
> $a[] =3D 'string'; // must be accepted
>
> $b =3D &$c->prop[1];
> $b[] =3D 'string'; // must be an error
> $c->prop =3D [];
> $a[] =3D 'string'; // must be accepted
>
> I don't remember all the possible cases, but I didn't find a way to
> support this that didn't involve recursively scanning an array at some
> point. IIRC, without references it's less of an issue, so a possible
> way forward would be to forbid references to members of typed
> properties. Unfortunately this breaks pass-by-reference, e.g.
> `sort($c->prop)`. out/inout parameters may be part of a solution, but
> with more array separations than pass-by-ref.
>
> Best Regards,
> Arnaud
>

Another one that I don't see mentioned that naturally follows from a
conversation I had with you a few weeks ago is operators on arrays. Namely,
the behavior of the `+` operator when used with arrays. How this would
interact with generics, and with different approaches to generics and
arrays, is probably something that will require attention. Operators in
general present some challenges (though not unsolvable ones, just
complicated ones) to languages that try to use both generics and loose
types, because operators generally don't have a way for the programmer to
help the engine with typing during the evaluation.

Jordan

--000000000000475b1206202a8ec2
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"ltr"><br></div><br><div class=3D"gmail_quote">=
<div dir=3D"ltr" class=3D"gmail_attr">On Tue, Aug 20, 2024 at 6:02=E2=80=AF=
AM Arnaud Le Blanc &lt;<a href=3D"mailto:arnaud.lb@gmail.com">arnaud.lb@gma=
il.com</a>&gt; wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"m=
argin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left=
:1ex">Hi Bob,<br>
<br>
On Tue, Aug 20, 2024 at 12:18=E2=80=AFAM Bob Weinand &lt;<a href=3D"mailto:=
bobwei9@hotmail.com" target=3D"_blank">bobwei9@hotmail.com</a>&gt; wrote:<b=
r>
&gt; The fluid Arrays section says &quot;A PoC has been implemented, but th=
e performance impact is still uncertain&quot;. Where may I find that PoC fo=
r my curiosity? I&#39;m imagining the implementation of the array types as =
a counted collection of types of the entries. But without the PoC I may onl=
y guess.<br>
<br>
I may publish the PoC at some point, but in the meantime here is a<br>
short description of how it&#39;s implemented:<br>
<br>
- The zend_array has a zend_type member representing the type of its elemen=
ts<br>
- Everytime we add or update a member, we union its type with the<br>
array type. For simple types it&#39;s just a |=3D operation. For arrays wit=
h<br>
a single class it&#39;s also simple. For complex types it&#39;s more expens=
ive<br>
currently, but it may be possible to cache transitions to make this<br>
cheaper.<br>
- Updating the array type on deletes requires to either maintain a<br>
counter of every type, or to re-compute the type entirely everytime.<br>
Both are probably too expensive. Instead, we don&#39;t update the type on<b=
r>
deletes, but we re-compute the type entirely when a type check fails.<br>
This is based on two hypotheses: 1. A delete rarely changes an array&#39;s<=
br>
type in practice, and 2. Type checks rarely fail<br>
- References are treated as mixed, so adding a reference to an array<br>
or taking a reference to an element changes its type to mixed. Passing<br>
an array&lt;mixed&gt; to a more specific array&lt;something&gt; will cause =
a<br>
re-compute, which also de-refs every reference.<br>
- Updating a nested element requires updating the type of every parent<br>
<br>
&gt; It also says &quot;Another issue is that [...] typed properties may no=
t be possible.&quot;. Why would that be the case? Essentially a typed prope=
rty would just be a static array, which you describe in the section right b=
elow.<br>
<br>
It becomes complicated when arrays contain references or nested<br>
arrays. Type constraints must be propagated to nested arrays, but also<br>
removed when an array is not reachable via a typed property anymore.<br>
<br>
E.g.<br>
<br>
class C {<br>
=C2=A0 =C2=A0 public array&lt;array&lt;int&gt;&gt; $prop;<br>
}<br>
<br>
$a =3D &amp;$c-&gt;prop[0];<br>
$a[] =3D &#39;string&#39;; // must be an error<br>
unset($c-&gt;prop[0]);<br>
$a[] =3D &#39;string&#39;; // must be accepted<br>
<br>
$b =3D &amp;$c-&gt;prop[1];<br>
$b[] =3D &#39;string&#39;; // must be an error<br>
$c-&gt;prop =3D [];<br>
$a[] =3D &#39;string&#39;; // must be accepted<br>
<br>
I don&#39;t remember all the possible cases, but I didn&#39;t find a way to=
<br>
support this that didn&#39;t involve recursively scanning an array at some<=
br>
point. IIRC, without references it&#39;s less of an issue, so a possible<br=
>
way forward would be to forbid references to members of typed<br>
properties. Unfortunately this breaks pass-by-reference, e.g.<br>
`sort($c-&gt;prop)`. out/inout parameters may be part of a solution, but<br=
>
with more array separations than pass-by-ref.<br>
<br>
Best Regards,<br>
Arnaud<br></blockquote><div><br></div><div>Another one that I don&#39;t see=
 mentioned that naturally follows from a conversation I had with you a few =
weeks ago is operators on arrays. Namely, the behavior of the `+` operator =
when used with arrays. How this would interact with generics, and with diff=
erent approaches to generics and arrays, is probably something that will re=
quire attention. Operators in general present some challenges (though not u=
nsolvable ones, just complicated ones) to languages that try to use both ge=
nerics and loose types, because operators generally don&#39;t have a way fo=
r the programmer to help the engine with typing during the evaluation.</div=
><div><br></div><div>Jordan<br></div></div></div>

--000000000000475b1206202a8ec2--