[prototype runtime] Use the original Module#name to avoid generating incorrect module name

[pocke]

Problem

rbs prototype runtime command generates RBS with incorrect module names if Module#name is re-defined.
For example:

# test.rb

module X
  module M
    def self.name() 'FakeNameM' end
    def self.to_s() 'FakeToS' end
    X = 1
  end

  class C
    def self.name() 'FakeNameC' end
    include M
  end

  class C2 < C
  end
end

$ exe/rbs prototype runtime -R test.rb X::*
class FakeNameC                 # ← It should be C, but not
  include FakeNameM             # ← It should be M, but not

  def self.name: () -> untyped
end

class FakeNameC < FakeNameC     # ← It should be X::C2 < C, but not
end

module X::M
  def self.name: () -> untyped

  def self.to_s: () -> untyped
end

FakeToS::X: Integer             # ← It should be X::M::X, but not

Solution

This pull request fixes this problem by using the original Module#name to get a constant name.
It will display the following result.

$ exe/rbs prototype runtime -R test.rb X::*
class X::C
  include M

  def self.name: () -> untyped
end

class X::C2 < X::C
end

module X::M
  def self.name: () -> untyped

  def self.to_s: () -> untyped
end

X::M::X: Integer

---

By the way, prototype runtime still uses other Module's methods that can be re-defined. For example, it uses Module#instance_method to get a method object of a module.
At first, I thought I should replace them with the original methods, but I didn't.
Because I guess it is not problematic in most cases. On the contrary, redefined methods can be more appropriate than the original methods.
For example, delegator's instance_method is re-defined to work with delegated methods. https://github.com/ruby/ruby/blob/53e352fd718cd2cae6d77298e6e92736dddcfeeb/lib/delegate.rb#L436
So I think we don't need to use the original method for them.

[soutaro]

👏