Looking at the your latest pr, it should be possible to write an abstract test for parMap/parTupled which works with n-ary implementations... But yeah having something n-ary is definitely preferred over copy pasta in tests as that becomes annoying if the functions do change at some point