Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:121566
Date: Fri, 03 Nov 2023 15:13:27 +0000
To: internals@lists.php.net
User-Agent: K-9 Mail for Android
In-Reply-To: <5ea314c4-f45d-4b7c-a94d-a82931eeb80b@daniil.it>
References: <CAJpLVh1P0kTneo+chFEdp2oCmSKNpgpDxSXd3atv9OHKejABgw@mail.gmail.com> <e5d9a346-015c-48d5-baa8-4742c78b600f@daniil.it> <CAPyj-LAfKEoOvaxMn1ggXcw4jaUfOzXEMAnMLydZq84tQGva6A@mail.gmail.com> <5ea314c4-f45d-4b7c-a94d-a82931eeb80b@daniil.it>
Message-ID: <EDD07186-FD62-4156-8BE6-6B4B714D1AD1@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable
Subject: Re: [PHP-DEV] Previous discussions about generics syntax only?
From: rowan.collins@gmail.com (Rowan Tommins)

On 3 November 2023 13:15:59 GMT, Daniil Gentili <daniil@daniil=2Eit> wrote:
>Well, I was actually thinking of enabling these compile-time checks only =
if preloading is enabled, to ensure the closed-world guarantee=2E

Doesn't that just take us back to the "generic types are only optionally c=
hecked" scenario you were trying to avoid?

Right now, I suspect the use of preloading is actually *less* common than =
the use of offline static analysers=2E Unlike OpCache itself, it's not some=
thing you can just switch on for an existing code base: you've got to switc=
h from rules for loading classes at runtime to generating a static preload =
list in advance=2E

To be reliable for static analysis, that preload list would need to be *co=
mprehensive*, defining every symbol up front, then making a second pass of =
cross-file analysis=2E That makes it look very much like a separate build s=
tep, rather than something integrated into the normal execution pipeline=2E=
 I've said before that this might be the way PHP should go - a native "modu=
le" concept, with full pre-compilation - but it would be quite a radical ch=
ange=2E

Perhaps an alternative is to change the rules of when autoloading can be i=
nvoked, by allowing the preload script to recursively autoload every symbol=
 it can identify, rather than waiting for the relevant code to be executed=
=2E I think this is how current offline static analysers work=2E

That doesn't solve the "optional checks" problem, though, if preloading re=
mains optional=2E We could produce an error if any generic/extended type wa=
s encountered without being preloaded, but that poses a dilemma for librari=
es: either don't use the new types, or impose a requirement for application=
s to use preloading=2E I'm not convinced there's an easy compromise here=2E

Regards,

--=20
Rowan Tommins
[IMSoP]