Provide duck-typed PSR-14 implementation

[weierophinney]

This patch provides a forwards-compatibility release that adapts zend-eventmanager to work as a PSR-14 EventDispatcher. It does so by doing the following:

• Creating package-specific ListenerProviderInterface and EventDispatcherInterface versions that work with PHP 5.6.
• Moving all listener aggregation into classes within a ListenerProvider subnamespace; this includes both the listener attachment previously in the EventManager instance as well as the SharedEventManager instance.
• Renaming "listener aggregates" to "listener subscribers" in the new ListenerProvider subnamespace.
• Adapting EventManager to consume listener providers. By default, it will create a default priority-based listener provider, and have its various listener attachment methods proxy to that provider; it then aggregates that provider with the SharedEventManager (which is itself a provider now) in order to retrieve listeners. This is done in such a way that behavior is completely backwards compatible with current usage.
  • A new named constructor, createUsingListenerProvider(), allows providing a specific listener provider for use with the EventManager. If the provider is attachment capable, the various listener attachment methods will proxy to it; otherwise, they will raise an exception.

The patch also deprecates a number of features, including:

• Most interfaces, including the EventInterface.
• The entire "shared event manager" concept (this can be accomplished via event object hierarchies instead)

TODO

• Write changelog entries
• Create documentation of new features
• Document how to start migrating to the new features in order to prepare for version 4

---

Originally posted by @weierophinney at zendframework/zend-eventmanager#73

[acelaya]

Hey!

Is there any plan to continue working on this? I see there was some work done on the old zendframework repo, but the pull request was not migrated to this one.

I'm asking because I plan to support both laminas-eventmanager and PSR-14 in a project of mine, and having this would drastically simplify things, haha.

I know there's a lot of work yet to be done in other packages, so I don't want to sound picky. I would just want to know if we are talking about a one month thing, a six month thing or a one year thing, just to decide if it's worth waiting or going with my own temporal implementation.

I'm also open to help with this if there's anything I can do.

[weierophinney]

@acelaya Well, clearly at this point, it's at least a six month thing. 🤣

I'd forgotten how much work I did on the original, and it's definitely a good starting place. I'll see if I can get some work done on it in the coming weeks, so we can get a v3 release with it in, and potentially prepare for a v4 release.

[weierophinney]

@acelaya See #12 - since 3.4.0 targets 7.3+, I was able to do a pretty straight implementation, instead of requiring duck-typing.

[acelaya]

Thanks for checking this now @weierophinney!