Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:124308
Precedence: bulk
MIME-Version: 1.0
References: <07A8534A-1B45-4F15-A78B-AA7DDF92B8B6@koalephant.com>
 <5BA404A4-7281-475F-A1F6-D1252ABB059D@miles.systems> <a69a759f-f185-42ff-b769-1f85beab673b@app.fastmail.com>
 <FD75AB78-4CF1-457B-8BFD-BE9F7AF723A9@miles.systems> <e867bbca-38c4-4523-916d-07e86bea94c9@app.fastmail.com>
 <39B8662A-FD7D-467A-883C-E8D71B5A4E02@miles.systems> <A800AB40-409B-4EC5-9B56-98968E707739@miles.systems>
 <CAAKU0unBzdf7UYr9xe21VxXyHybvVZx9_Fc9DfP-i-0fYoddEw@mail.gmail.com> <09B5B065-35B2-4E67-9EDA-DF6076F4B2ED@miles.systems>
In-Reply-To: <09B5B065-35B2-4E67-9EDA-DF6076F4B2ED@miles.systems>
Date: Tue, 9 Jul 2024 10:53:42 -0600
Message-ID: <CAAKU0u=wmHe04YFk8tSW-Dhw5d4+za5RUAG=UVZy6Q_WEi6tMw@mail.gmail.com>
Subject: Re: [PHP-DEV] [Initial Feedback] Typed Arrays
To: Richard Miles <richard@miles.systems>
Cc: php internals <internals@lists.php.net>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
From: levi.morrison@datadoghq.com (Levi Morrison)

On Tue, Jul 9, 2024 at 10:18=E2=80=AFAM Richard Miles <richard@miles.system=
s> wrote:
>
> Hi Levi,
>
> First, were you able to discuss this with authors of other generic
> proposals or implementations? Derick Rethans and Arnaud Le Blanc both
> have experience there.
>
>
> This proposal has nothing to do with generics.
> Derick Rethans responded in this thread, and I've specifically addressed =
his current (generic?) branch.
> It is a community misconception that typed arrays have anything to do wit=
h Generics and also
> has been discussed in this thread in detail.
> Is gatekeeping you asking me to repeat what was discussed back to you
> specifically so you can grant me privileges? I digress and I'll concede.

Please calm down. I am not gatekeeping, as I don't even have the
know-how to grant you the wiki karma.

> Second, do you have any work towards an implementation? Technically,
> an implementation is not required for RFCs. However, for this feature,
>
>
> I did start on the implementation and have linked the branch above, but s=
topped after
> Rob Landers=E2=80=99 suggestion:
>
> In my honest opinion, before we can even have this type of conversation, =
we need to have an RFC about the syntax (there may already be one, I haven=
=E2=80=99t checked =E2=80=94 and internet is spotty where I currently am). =
The pattern matching RFC was proposed as though it would be written one way=
, people have thrown out ideas of different syntax, but I don=E2=80=99t thi=
nk there is an =E2=80=9Cofficial=E2=80=9D syntax. A good starting point may=
 be to simply propose an RFC about syntax so future RFCs can be consistent.
>
> =E2=80=94 Rob
>
>
>
> Also, directly from the how-to page on the mailing list "https://wiki.php=
.net/rfc/howto"
>
> "Email internals@lists.php.net to measure reaction to your intended propo=
sal.
> State who would implement the feature, or whether the proposal is only a =
=E2=80=9Cconcept=E2=80=9D.
> Proceed with an RFC if feedback is not negative or if a detailed RFC will=
 clarify the proposal. =E2=80=9C
>
> So, unless I'm mistaken, I think we=E2=80=99re going about this correctly=
.? I=E2=80=99m not sure whoes 'voice to follow=E2=80=99 since
> Rob and Levi have differing opinions on paths forward. I personally would=
 like the syntactic approval before
> I implement the full code. The complete implementation details would need=
 its own discussion.

Syntax is part of the implementation, but it's not the hard part from
a technical perspective. The hard part is actually implementing the
semantics and behavior, and plumbing it through all the parts of the
engine. It does not matter what the "skin" of the thing looks like, if
the mechanics of the things simply don't work.

Also, I don't mean to say you need a complete implementation, either.
More of a "proof of concept" which means that core issues have been
identified, and are either fixed or outlined so other people are aware
of them.

> I can guarantee any RFC without an implementation is dead-on-arrival.
>
>
> You're not the first person to have a guarantee in this thread=E2=80=A6 :=
)

No, but I authored and co-authored many RFCs over many years. I was
trying to be helpful. I'll state that I don't think your RFC has any
chance of success. The syntax, the semantics, the discussion so far,
especially that typed arrays have nothing to do with generics... I
think its chance of success is quite low. I messaged you to try and be
helpful despite my own opinions on such things, and tried to point you
to things that I thought would be helpful if you wish to pursue this:

 1. A proper discussion with authors of other proposals, including
ones related to generics, is prudent. Even if you believe they aren't
related, I'm sure they've learned things about the engine that would
be helpful to you in your implementation.
 2. A proposal for a significant feature without the bones of an
implementation simply will not pass. Since the RFC process was
created, I don't think it's ever happened. There have been cases where
the implementation changed from RFC to what landed (Dmitry completely
reimplemented from scratch one of my RFCs that passed, I think it was
return types), but a basic implementation is needed to fully
understand the problems and challenges of the implementation.

Lastly, I'll note that I responded to you off-list. I don't know why
you brought a private conversation back to the list without discussing
it with me. There wasn't anything rude or inappropriate, and the rest
of the list certainly could have gone without this conversation. We
should respect the time of everyone participating on the list.