Metamodel improvement: introduce Module

[jhass]

I think we should further copy the meta model from Ruby, even if it's just fake.

Code	Current	Expected
Module	Undefined constant	Module
Module.class	"	Class
Module.is_a? Object	"	true
class C; end; C.is_a? Module	"	true
module M; end; M.is_a? Module	"	true
class C; end; C.is_a? Class	true	true
module M; end; M.is_a? Class	true	false
module M; end; M.class	Class	Module
module M; end; M.class.class	Class	Class

[technorama]

+1

[sdogruyol]

This is pretty sweet 👍

[PragTob]

Looks good to me 👍

[asterite]

We'll probably do this with time, although I always wonder what's the use for a metamodel, and specially the need to distinguish between a module and a class at runtime.

[jhass]

It's mostly introspection. Sometimes it's just quicker to do p Something.class than consulting documentation. The other point is consistency, there's a certain beauty to the root of the type system following the same rules than your own stuff.

[bjmllr]

Should Crystal's Module have comparison operators (< <= <=> > >=) like those provided in Ruby's Module to query the variance between two modules? Also related to #1297

[sardaukar]

👍

[kuraga]

👍

[asterite]

You can add emojis to the issues themselves, there's no need to add 👍 comments anymore