Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:129864
Feedback-ID: i8414410d:Fastmail
Precedence: list
MIME-Version: 1.0
Date: Thu, 22 Jan 2026 10:35:53 -0600
To: "php internals" <internals@lists.php.net>
Message-ID: <91952aa8-5d1e-401d-9f43-c812036e358f@app.fastmail.com>
In-Reply-To: <28461544eb11b41dce7c7563ad879e63@bastelstu.be>
References: <64953ec741a4c6609519e1878ad37b54@bastelstu.be>
 <253ee811-0c1d-7eab-4c36-38607401aaf6@php.net>
 <28461544eb11b41dce7c7563ad879e63@bastelstu.be>
Subject: Re: [PHP-DEV] [RFC] Partial Function Application for instance of non-static
 methods ("$this")
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
From: larry@garfieldtech.com ("Larry Garfield")

On Thu, Jan 22, 2026, at 10:08 AM, Tim D=C3=BCsterhus wrote:
> Hi
>
> Am 2026-01-22 16:33, schrieb Derick Rethans:
>>> https://wiki.php.net/rfc/partial_function_application_this
>>=20
>> 	We propose allowing a named parameter $this: ? to be combined with t=
he
>> 	=E2=80=9Cscope resolution operator=E2=80=9D syntax of referring to a=
 method.
>>=20
>> I can't say I am entirely sold on the syntax here =E2=80=94 starting =
with a $
>> makes it look like a variable.
>
> Yes, it certainly requires some getting used to it. It's the best we=20
> could come up with, but if there are any better suggestions, we're ope=
n=20
> to hearing those. I'd like to note the =E2=80=9CRejected Features=E2=80=
=9D section,=20
> since it lists some non-workable alternatives we considered.

I too am very much in favor of this functionality, though not a fan of t=
he syntax either.

My previous, un-finished thinking was for `$$->foo('blah')`, which would=
 translate to a mixed fn that just forwards the call.  I... hadn't actua=
lly thought about also partially applying the method, since the use case=
s I can think of would all be "call this method on the value piped from =
the previous pipe step."  (Or, read this property.)  That would allow it=
 to be embedded within an array_map() call or similar, since it's techni=
cally independent of both PFA and pipes.

If it's being partially applied, then I don't see a way around the redun=
dant typing, which is annoying.  At least not without vastly better type=
 inference than we have today.

So, I think I am mainly interested in the degenerate case where $this is=
 the only thing being partially applied, which seems like a common enoug=
h case that it could have a nicer version?  I'm just spitballing at this=
 point, I think.  But I do like the direction.

--Larry Garfield