Optimized version of Tuple#to_a

[asterite]

Benchmark:

require "benchmark"

Benchmark.ips do |x|
  x.report("old") do
    {1, 2, 3, 4}.to_a
  end
  x.report("new") do
    {1, 2, 3, 4}.to_a2
  end
end

Results:

old  24.58M ( 40.68ns) (± 2.92%)  64.0B/op   1.42× slower
new  34.95M ( 28.62ns) (± 2.73%)  64.0B/op        fastest

It's not much but it's something.