Add tests to check that the "unicode61" and "icu" tokenizers both identify white-space codepoints outside the ASCII range.

FossilOrigin-Name: bfb2d4730cbbe18fb940e72f4fde9122d550734e