Upgrade Rubocop/Erblint and fix cop violations

[murny]

Replaces #1749

Upgrade rubocop and erblint gems to latest versions

Fix code to make new cops happy, mainly Style/ArrayCoercion and Style/StringConcatenation

[murny]

This cop wants us to avoid using .insert because it skips validations. But it's a false positive as we using .insert from our rdf graph object instead of ActiveRecord.

app/controllers/aip/v1/collections_controller.rb:24:31: C: Rails/SkipsModelValidations: 
Avoid using insert because it skips validations. 
(https://guides.rubyonrails.org/active_record_validations.html#skipping-validations)

      rdf_graph_creator.graph.insert(RDF::Statement(statement_definition))
      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

So just disable this cop from these files

[murny]

Rubocop prefers string interpolation over string concatenation. Updates all our violations of this.

app/decorators/facets/member_of_paths.rb:10:26: C: 
Style/StringConcatenation: Prefer string interpolation to string concatenation. 
(https://rubystyle.guide#string-interpolation)
                         ‘/’ + Collection.find(collection_id).title
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

[murny]

app/models/exporters/solr/base_exporter.rb:183:7: C: 
Style/ArrayCoercion: Use Array(role) instead of explicit Array check. 
(https://rubystyle.guide#array-coercion)
      role = [role] unless role.is_a? Array
      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Interesting new cop. Essentially it argues you should Use Array() instead of explicit Array check or [*var], when dealing with a variable you want to treat as an Array, but you’re not certain it’s an array.

So switched any violations over to new syntax

[mbarnett]

Going to suggest that we disable this cop/revert these changes (I vaguely think there might have been an older cop we disabled for the same reason?). Array() is kind of a subtle foot-gun, and I don't think Batsov understands that...

The difference is this:

irb(main):001:0> a = {:a => 1, :b => 2}
=> {:a=>1, :b=>2}
irb(main):002:0> array = [a] unless a.is_a? Array
=> [{:a=>1, :b=>2}]
irb(main):003:0> different_array = Array(a)
=> [[:a, 1], [:b, 2]]

Instead of putting a hash in an array, it blows it up into keywise paired subarrays, which generally isn't what people expect, making it not semantically equivalent to the [x] unless x.is_a? Array pattern.

[murny]

Makes sense. We can play it safe here and revert these 👍

[mbarnett]

Going to suggest that we disable this cop/revert these changes (I vaguely think there might have been an older cop we disabled for the same reason?). Array() is kind of a subtle foot-gun, and I don't think Batsov understands that...

The difference is this:

irb(main):001:0> a = {:a => 1, :b => 2}
=> {:a=>1, :b=>2}
irb(main):002:0> array = [a] unless a.is_a? Array
=> [{:a=>1, :b=>2}]
irb(main):003:0> different_array = Array(a)
=> [[:a, 1], [:b, 2]]

Instead of putting a hash in an array, it blows it up into keywise paired subarrays, which generally isn't what people expect, making it not semantically equivalent to the [x] unless x.is_a? Array pattern.

[pgwillia]