I'm too eager at merging today :D

@ondrejmirtes Good for me 😆 I'll create another pull request for phpstan to remove the parameter and ignore any remaining errors.

So, apparently there's one more bug when resolving phpDocs from a method in a nested trait. I'll take a look at that tomorrow.

BTW: 90% real-world problems at my job's codebase the rule found (with reportMaybes false) is about overriding parent return type void with something specific. I think we could relax this a bit, because I think it doesn't break LSP - when the parent interface is typehinted, the user code doesn't read the result, so it isn't bothered when a child implementation returns something there...

@ondrejmirtes True. I can adjust that tomorrow.

One more thing I'm realizing - the rule should probably be switched off for static methods. Yes, it can be invoked like $foo::doSomething() and LSP applies there, but that's pretty uncommon.

I did the proposed changes on phpstan/phpstan master, please check them out :)

Nice, thank you 🙂