Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:116719
MIME-Version: 1.0
References: <CAL8UiQXS_QbDo33KRX_SXbR3jsDOzG2Q14kUpq1CuWsyQpHyZA@mail.gmail.com>
 <CAAKU0ukwH9+KXkfY2zpm3OWckN44LhLRNSN4cE=bz5uNGXyQ-w@mail.gmail.com> <CAL8UiQXkMV8zRPeKR8m0AY8vfamCJ_1b9qz+JefeSQoLqMCZsQ@mail.gmail.com>
In-Reply-To: <CAL8UiQXkMV8zRPeKR8m0AY8vfamCJ_1b9qz+JefeSQoLqMCZsQ@mail.gmail.com>
Date: Wed, 22 Dec 2021 07:51:39 -0800
Message-ID: <CAMrTa2GPoL_uWO8rv-1J9fw8zPJkW9-xWExBj3zXazcRfmjjsw@mail.gmail.com>
To: Hassan Ahmed <7snovic@gmail.com>
Cc: Levi Morrison <levi.morrison@datadoghq.com>, php internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary="0000000000009d4c6e05d3be1954"
Subject: Re: [PHP-DEV] [VOTE] Add array_group function
From: jordan.ledoux@gmail.com (Jordan LeDoux)

--0000000000009d4c6e05d3be1954
Content-Type: text/plain; charset="UTF-8"

On Wed, Dec 22, 2021 at 1:39 AM Hassan Ahmed <7snovic@gmail.com> wrote:

> Hello Levi, maybe, I just was confused while I was rewriting it. any tips
> would be appreciated.
>

There's a lot of information about this that is missing from the RFC. It
says it "will be overwritten". Does that mean that it does a right union?
What exactly is the meaning of null for the parameters? That isn't covered.

The language on how it handles key conflicts is confusing. It says that it
"currently" handles it by "overwriting", but doesn't explain what that
means. It also doesn't *currently* do that, since it isn't currently part
of the language. You have to scroll down to the example to decipher that
"overwrite" means it takes the last value.

It says there's no backwards incompatible breaks, but that isn't strictly
true, as the `array_group()` function in the global namespace will now be
occupied. That *might* be a rather small BC break, but I'm not sure.

I think you are seeing no votes mostly because the RFC itself is simply
incomplete. It doesn't describe the use case, it doesn't describe the
implementation at all, and it doesn't describe the edge cases at all. I
think it could be an idea that people might support, but you probably need
to find someone who has done an RFC before to help you workshop it a bit.
These are all fixable problems in the RFC document, it could be that the
implementation and the idea are well received.

Jordan

--0000000000009d4c6e05d3be1954--