Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:109007
User-Agent: Cyrus-JMAP/3.1.7-991-g5a577d3-fmstable-20200305v3
Mime-Version: 1.0
Message-ID: <a46d0ba9-2981-4c58-afb1-f648c499450e@www.fastmail.com>
In-Reply-To: <102D234E-BC5D-42B2-ABC4-93980C1F76A4@pmjones.io>
References: <50BD013E-CF72-414C-BBC0-A7A2E45CBDDB@pmjones.io>
 <D2321EFF-6D87-422A-BCC3-875820ADF8E2@pmjones.io>
 <C2A122BA-052B-4AFB-900E-93858BF50F7E@koalephant.com>
 <0B40E6E5-342F-4D81-9CAA-A0C0739A7718@pmjones.io>
 <C3C15896-923D-4386-A991-329D0FCE7CC9@koalephant.com>
 <102D234E-BC5D-42B2-ABC4-93980C1F76A4@pmjones.io>
Date: Fri, 13 Mar 2020 15:36:04 -0500
To: "php internals" <internals@lists.php.net>
Content-Type: text/plain;charset=utf-8
Content-Transfer-Encoding: quoted-printable
Subject: Re: [PHP-DEV] RFC: Server-Side Request and Response Objects (v2)
From: larry@garfieldtech.com ("Larry Garfield")

On Fri, Mar 13, 2020, at 9:55 AM, Paul M. Jones wrote:

> > As with anything any of us has written, I=E2=80=99m not 100% sold on=
 =E2=80=98Current{Request,Response}` even after writing it, but I think =
it=E2=80=99s at least a little more specific about what they do, when th=
e namespace is taken into account.
>=20
> `Current{...}` is not something we had previously considered; that=20
> would net us, in the global namespace:
>=20
> - CurrentRequest
> - CurrentResponse
> - CurrentResponseSender
>=20
> I will need to ponder on those as well.
>=20
> Any further thoughts or opinions on this, Stephen? Or from anyone else=
?

I am still negative on the RFC overall, but I'll throw this out there:

* The RFC by design is trying to take the super-globals and make them OO=
Py; no more, no less.
* The super-globals are not based on HTTP.  They're based on CGI, which =
is sort of but not quite an HTTP request.
* So... call it CgiRequest/CGIRequest?  (I am also ignoring capitalizati=
on.)  Because that's what it is: It's CGI request data wrapped up into a=
n object, no more, no less.

(The fact that it is no more is the reason I'm not a fan, but at least t=
hen it's accurate as to what it is, and isn't, and doesn't namespace clo=
bber those libraries that are modeling HTTP, rather than CGI.)

The response object is less CGI-bound, so it could still be ServerRespon=
se since it's a response from a server, and not intended for any other u=
ses (like HTTP client libraries). =20

--Larry Garfield