Make sure to avoid Surrogate Code points

[klnusbaum]

From the Rust documentation:

A char is a ‘Unicode scalar value’, which is any ‘Unicode code point’ other than a surrogate code point. This has a fixed numerical definition: code points are in the range 0 to 0x10FFFF, inclusive. Surrogate code points, used by UTF-16, are in the range 0xD800 to 0xDFFF.

No char may be constructed, whether as a literal or at runtime, that is not a Unicode scalar value

We should do something like "for any generated codepoint equal to or above this range, shift it up by the size of the range so we can guarantee we never put anything in the range, but can also shift back down when decrypting"