[3.x] Lazy services

[voronkovich]

Pull Request for Issue joomla/joomla-cms#45062

This PR add an ability to define lazy services using a new PHP 8.4 feature: https://www.php.net/manual/en/language.oop5.lazy-objects.php

Summary of Changes

A new shortcut method lazy() was added:

$container->lazy(
    Foo:class,
    function($container) {
        return new Foo($container->get('bar'));
    }
);

// Proxy object
$foo = $conainer->get(Foo::class);

// Factory is called here
$foo->someMethod(); 

Testing Instructions

Unit tests are provided.

Documentation Changes Required

Yes.

[HLeithner]

I don't really see the benefit of knowing if a object is lazy, if the container or a consumer needs to know it (what I'm not sure if it is needed) it could be achieved with a reflection and the isUninitializedLazyObject() method.

What would be the use case for this implementation?

[voronkovich]

About use case for this feature you can read here: joomla/joomla-cms#45062

This allows to instantiate the plugin only when event is triggered, and saving some resources (time and memory).

We can achieve it by defining plugins as lazy.

Also you can read about lazy services here: https://symfony.com/doc/current/service_container/lazy_services.html

[voronkovich]

The ContainerResource class is final and it's marked as internal, so changing it wouldn't cause the BC break.

[laoneo]

To simplify the whole pr, I would only add a lazy function and then create an insteance of resource there. The rest can be reverted I guess as it complicates the whole thing. Also in the resource I would only add the really needed functions in a first implementation. If we need more along the way we can always add them when needed.

[laoneo]

But I like that you added unit tests to the pr.

[voronkovich]

@laoneo, I've added the changes you requested. But I'm not sure if I got them right.

[laoneo]

That's exactly how I intended it.

[laoneo]

I made a more lightweight version in #63. Can you have a look if I missed something?

[voronkovich]

Closed in favor of #63