Thanks! I corrected the property using NonNegative.

Relying blindly on Quickcheck is bad practice. For instance in this case, by removing the newtype Input definition, the random tests now generate arrays with negative numbers, which goes against the description.

Thank you very much for your contribution! I tried to simplify the tests in my fork of your translation.

isnt that slow?

gnihihi

black magic

copious copies of liftAN and compose really can make absolutely anything pointfree

Thanks!

Fixed

That's awesome, thanks!

This appears to only be an issue with the sample tests (the full tests have gen15 = choose (0, 15)). My recently published fork fixes this in the sample tests.

!-)

Seems to be the oldest thread here from ten years ago, but I don't see it resolved in the code.

Please correct me if I'm mistaken:

The maximum degree of our polynoms will be smaller than 16.

In the Haskell property test suite though there is gen15 = choose (0, 16) which generates polynoms x^16 as the range of choose is inclusive.