Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:123275
Message-ID: <59d9c213-e159-420f-821c-46a6325f943d@bastelstu.be>
Date: Tue, 7 May 2024 20:32:29 +0200
Precedence: bulk
MIME-Version: 1.0
Subject: Re: [PHP-DEV] RFC [Discussion]: array_find
To: =?UTF-8?Q?Joshua_R=C3=BCsweg?= <josh@wcflabs.de>,
 PHP internals <internals@lists.php.net>
References: <24e4529d-0b75-44de-90ef-34de5dfb1c99@wcflabs.de>
 <278889be-82ab-4827-a9e7-801b5ba2d8f8@app.fastmail.com>
 <CAAKU0ukwfF+D5L+23BEqL053ktCp5AQOjRGdha-8pXN-yf1HBQ@mail.gmail.com>
 <06373f2b-5de0-4582-96c5-29c3b474c01d@wcflabs.de>
 <45e6365b-4963-4969-8cc1-80bae6922fc8@wcflabs.de>
 <60bb55a5-16c9-4df0-8521-75f82d3dbf5b@app.fastmail.com>
 <41ad56d9-8b27-420e-ba85-a3e9a6122bcf@wcflabs.de>
 <7a77288e-ddb7-436e-8981-b8b2508e1547@app.fastmail.com>
 <49aa8581-ed2e-4aa3-b330-7cd5d8efa764@wcflabs.de>
Content-Language: en-US
In-Reply-To: <49aa8581-ed2e-4aa3-b330-7cd5d8efa764@wcflabs.de>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
From: tim@bastelstu.be (=?UTF-8?Q?Tim_D=C3=BCsterhus?=)

Hi

On 5/6/24 20:44, Joshua Rüsweg wrote:
> I don't think that any internal functions are realistically passed
> there, but rather user-defined ones. I would therefore not make the
> function any more complicated than it already is. If the function works
> differently with internal functions than with user functions, this can
> also be very confusing for users.
> 
> [...]
To have another non-RFC-author opinion on the record: I agree here. 
Making the proposed functions detect the signature of the passed 
callback would *add* to the inconsistency, rather than removing from it. 
They would be the only functions doing that.

I'd rather see this fixed properly by making the behavior of internal 
and userland functions consistent when facing superfluous parameters 
(with a preference for the stricter behavior of internal functions), as 
has been discussed in the ML thread that Joshua linked.

I'm in favor of the RFC and implementation the way it currently looks 
like. Implementation just needs some additional test cases, e.g. to 
ensure the callback is not called after an item has been found and to 
ensure proper behavior when an Exception is thrown.

Best regards
Tim Düsterhus