I'm writing a bundle for GNU Forth - mainly for giggles although I'll release it if it gets good enough. I'd like ⎋ completion to be able to complete Forth words - which can include any non-space character. Is it possible to define the 'word characters' on a per-grammar basis?