Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:124761
Message-ID: <519bd3144e5b99a57d5d411a99dcdc0077261c33.camel@ageofdream.com>
Subject: Re: [PHP-DEV] [RFC] Add Directive to Make All Namespaced Function
 Calls Global
To: internals@lists.php.net
Date: Mon, 05 Aug 2024 08:01:35 -0400
In-Reply-To: <0e04da48-0d16-f5d5-29b3-b43271d0d0d2@php.net>
References: <d9c62789e46d1f3bbb67b9c966bb613661f7889b.camel@ageofdream.com>
	 <0e04da48-0d16-f5d5-29b3-b43271d0d0d2@php.net>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.46.4-2 
Precedence: bulk
MIME-Version: 1.0
From: lists@ageofdream.com (Nick Lockheart)

On Mon, 2024-08-05 at 12:27 +0100, Derick Rethans wrote:
> On Sun, 4 Aug 2024, Nick Lockheart wrote:
>=20
> > https://wiki.php.net/rfc/global_function_parser_directive=C2=A0I am
> > asking=20
> > that we discuss and vote on the following question:
> >=20
> > =E2=80=9CShould there be some way for developers to signal to the parse=
r at
> > compile time that all unqualified function names found in a
> > namespace=20
> > context are global, without a namespace lookup?=E2=80=9D
>=20
> I don't beleive that we should introduce a switch for this at all,
> just=20
> like the conclusion was in
> https://wiki.php.net/rfc/use_global_elements
>=20
>=20

In that RFC, the proposal was:

declare(function_and_const_lookup=3D'global');

I'm not sure the conclusion was that there shouldn't be a way to
control the NS lookup behavior. That's what I'm trying to determine
with this new RFC: whether it was the entire concept, or just the
syntax or mechanism that was disfavored.

Do people agree that there should be=C2=A0*some way* to use only global
functions without an NS lookup as a *concept*?

Or, did they not like the syntax of:

declare(function_and_const_lookup=3D'global');

?


Because that RFC has four things in it:

(1) A Concept that global functions should be used without an NS lookup
(2) A Concept that global *constants* should be usable without an NS
lookup.
(3) A syntax being proposed.
(4) The idea that functions and constants should both be controlled by
the same directive.


I think that other RFC was on the right page, for the most part, but I
think the syntax could be improved.

Which is why I had suggested a new keyword as part of the namespace
declaration in some of my other posts.

Something that would look like one of these:

// Disable NS lookup; use all *global* functions unless
// a \foo\ appears before the function name:

namespace foo using global functions;


// Disable NS lookup; use all *local* functions unless
// a \ appears before the function name:

namespace foo using local functions;


// Use NS lookup. Same as today, no BC break:
namespace foo;=20