I found that the Arbitrary instance for the Char type generates only ASCII values while Char type represent Unicode characters.

I can't figure out why it has such a limit. Not knowing this pitfall, one can misjudge the test results.

Is this intended thing or just not being implemented yet?