[5.4] CMSPlugin: deprecation for registerListeners

[Fedik]

Pull Request for Issue # .

Summary of Changes

The PR deprecate CMSPlugin::registerListeners() as no longer needed when plugin will implement SubscriberInterface
(I wonder why it was not marked as deprecated before).

Testing Instructions

Apply PR, and navigate around the site, all should work as before.

The PR requires review from maintainers.

Link to documentations

Please select:

• Documentation link for docs.joomla.org:
• No documentation changes for docs.joomla.org needed
• Pull Request link for manual.joomla.org: CMSPlugin: deprecation for registerListeners Manual#422
• No documentation changes for manual.joomla.org needed

Ping @HLeithner @wilsonge

[HLeithner]

make sense for me

[laoneo]

I guess it was not marked as deprecated before becaus we changed the plugins way later.

[Fedik]

Ah, wait, I wanted to write different, not in 5.4 but in 6.0.
This:

• in 5.4 for plugins with PluginWithSubscriberInterface (new in 5.4) we ignore registerListeners() and DispatcherAwareInterface (as it done in this PR)
• in 6.0 we remove DispatcherAwareInterface interface from CMSPlugin, but will keep traits. And also add check for DispatcherAwareInterface for plugins which explicitly use it on their class (similar to as it done in [6.1] CMSPlugin: use Lazy Object feature #45062 but for all plugins)
• in 7.0 we remove traits

[HLeithner]

Introducing a new Interface makes no sense to me, just remove the DispatcherAwareInterface in 6.0 and the trait in 7.0.

Only call the registerListeners() if SubscriberInterface is not set, maybe a bit more b/c friendly could be to check if the the method is overwritten and call it if it's overwritten but to be honest I would do that.

[Fedik]

Only call the registerListeners() if SubscriberInterface is not set

That is not possible, we already discussed that. It will be b/k break.

[HLeithner]

yes I know, but not if you check if the function is overridden or?

[Fedik]

Maybe, but checking whether function is overridden would require use of reflection for every plugin. Use of interface is much cleaner.
Or I missing something?

[HLeithner]

Yeah the interface might be cleaner, but it also means a 3rd party dev needs to touch the plugin for this reason 2 times...

An alternativ could be to override the methods in with dummy/empty methods in core this would not initialise the object and wouldn't need to be removed (even if it should) later to be b/c.

especially if you like that your plugin works between 4.0 and 7.4 it's hard (not really possible with out defining your own poly-fill) to use a new interface,

[laoneo]

Agree here with @HLeithner, as removing the interface will be problematic. I keep my extensions at least the two major version compatible, when not even three. So having a required interface in 6 which get removed in 8 and is still used in 7 would be problematic.

[laoneo]

I think we need to accept that plugins with DispatcherAwareInterface cannot be lazy.

I'm fine with this. In the provider.php the plugin knows if the setDispatcher function exists or not to inject the dispatcher.

[Fedik]

I removed the extra interface from the PR.
Will find a solution in another PR, here just deprecation.

[laoneo]

I have tested this item ✅ successfully on bff97f4

Deprecation messages do look ok

---

_{This comment was created with the J!Tracker Application at issues.joomla.org/tracker/joomla-cms/43395.}

[richard67]

The PR for the manual joomla/Manual#422 is made for the migrations/52-53/new-deprecations.md document. Should that not be migrations/53-54/new-deprecations.md?

[richard67]

I've restored @laoneo 's test result in the issue tracker as the commit which invalidated the test count was just a grammar change in a comment.

[richard67]

I have tested this item ✅ successfully on bb9a233

---

_{This comment was created with the J!Tracker Application at issues.joomla.org/tracker/joomla-cms/43395.}

[richard67]

RTC

---

_{This comment was created with the J!Tracker Application at issues.joomla.org/tracker/joomla-cms/43395.}

[HLeithner]

thanks all for participation