prelude: | str = "a" * 1000 re = Regexp.new(str) benchmark: dup: re.dup