Document encoding in the manual

[glebm]

Things like:

1. The .waxeye grammar file is encoded as UTF-8.
2. The generated runtime files are encoded as UTF-8.
3. All runtimes use their native string types (C uses char *). Positions reported by the parser are offsets indices of the native string type.
4. The parser operates on Unicode codepoints, parser input must be well-formed (decodable to a sequence of Unicode codepoints).
5. Waxeye does not handle Unicode normalization. If you care about this, normalize (e.g. NFC-normalize) both the grammar file and the parser inputs.
6. Case-insensitive literals are ASCII-only.

Parser runtimes that support Unicode: C (TBD), JavaScript (#47), Ruby, Racket.