Re: Re: RFC: Nested Classes

From: Date: Wed, 07 May 2025 13:18:54 +0000
Subject: Re: Re: RFC: Nested Classes
References: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18  Groups: php.internals 
Request: Send a blank email to [email protected] to get a copy of this message
Hi

Am 2025-05-06 22:04, schrieb Rob Landers:
I think these are fundamental problems (if they are a problem at all) with how PHP currently does namespaces and names.
I don't think that this is a fundamental problem of namespaces and names. Ilija solved the naming conflict issue in his file-private class proposal.
I'm curious if you would vote "no" for namespace visibility as well?
I cannot answer this question. PHP doesn't have namespace visibility and whether or not I would be in favor of a proposal depends greatly on the proposed semantics. In the current nested classes RFC, the proposed semantics for private classes are inconsistent with how private in PHP currently behaves, which makes it unacceptable to me.
For everyone else who voted "no"; I would sincerely love to know the reason. Even if you just email me directly; it would really help me understand what can be improved in any future RFCs.
I unfortunately didn't have the time to follow the discussion and think about the RFC in depth. Besides the private semantics, the autoloading impact is also a no for me, since to be able to properly use nested classes, a composer update would be required, so that composer’s autoloading implementation becomes aware of nested classes to properly load them. This means that if a package uses nested classes, it will *silently* be broken until the *consumer* updates their composer installation. I don't think it is currently possible to define a minimum composer version as part of a package’s dependencies. Best regards TIm Düsterhus

Thread (102 messages)

« previous php.internals (#127300) next »