ConstantName should allow CamelCase constants for class aliases

[marcandre]

The follow code is flagged by RuboCop:

class Foo
end
Bar = Foo

I believe it shouldn't complain about Bar being camelcase here. Assuming Foo is correctly written, then Bar should actually be CamelCase, and BAR = Foo should actually be an error.

I understand there is rarely a case for such code (but we have nevertheless 14 such cases 😈 )

$ rubocop -V
0.51.0 (using Parser 2.4.0.0, running on ruby 2.3.5 x86_64-darwin15)

[garettarrowood]

This sounds like a fun fix. Happy to take a stab at it over the next few days.

---

Looks like there is already a note about this in the codebase - https://github.com/bbatsov/rubocop/blob/master/lib/rubocop/cop/naming/constant_name.rb#L27-L30

[garettarrowood]

As that comment suggestions, it looks like the described behavior actually is not flagged by rubocop. I tried to reproduce without success. Adding this spec also passes:

  # spec/rubocop/cop/naming/constant_name_spec.rb
  it 'does not throw offense if rhs is camelCase' do
    expect_no_offenses(<<-RUBY.strip_indent)
      class FooBaz
      end
      BarBaz = FooBaz
    RUBY
  end

@marcandre - Can you provide any additional info about the error?

[marcandre]

Indeed, Foo = Bar is ok, but Foo = Bar = Qux is flagged.

[marcandre]

Thanks for the fix. On the other hand, this allows Foo = BAR and Bar = FOO, right?

[garettarrowood]

From a quick test, it looks like those examples work just fine. No offenses on my end.

[marcandre]

Yeah. Maybe it should be an offense though, at least Foo = BAR. Not sure if FOO = Bar is ok.