Add class_implements return type extension

[herndlm]

Refs: phpstan/phpstan#8059
Closes phpstan/phpstan#4335

See also https://3v4l.org/IufYX

If this is OK, I can do the same for the other related functions either here or in dedicated PRs.

[herndlm]

The one nextras failure should be fine. They do a class_exists before which makes the string a class-string and therefore false will never be returned: https://github.com/nextras/orm/blob/v4.0.6/src/Entity/Reflection/MetadataParser.php#L421

[ondrejmirtes]

I don't think this is worth it. You can say these things for sure only for final classes, otherwise the array isn't really complete - a subclass can implement more interfaces.

[herndlm]

I don't think this is worth it. You can say these things for sure only for final classes, otherwise the array isn't really complete - a subclass can implement more interfaces.

It should work fine with subclasses as well, why wouldn't it? Or do you mean something different?

[herndlm]

aaah I think I got you now, makes sense

I'll have to find another fix for the webmozart assert problem then :)

[VincentLanglet]

A return type extension is still useful for phpstan/phpstan#4335 @herndlm
Instead of looking for the exact values returned by class_implements just knowing if the return value is class-string[] or class-string[]|false would be great.

Do you want to reopen the PR or should I work on it ? :)

[herndlm]

You can go ahead, I fixed my problem that I was having in a different way :)

[VincentLanglet]

You can go ahead, I fixed my problem that I was having in a different way :)

I started #2023 then if you have time to take a look.
You recently introduce a lot of new method in the Type interface, and I may miss something but didn't find the one I needed and had to use instanceof.