Use case: `InputRef::skip_n` or equivalent about chumsky

[safasofuoglu]

lexical-numbers on main hits the spot for me.

I could probably implement a Input::skip_n method that's exposed via InputRef with a default unoptimised implementation that just called next_maybe repeatedly though. Would this suit your needs?

That's not any better than the two approaches I tried above. I was thinking about a specialized AnyRepeated that eats n bytes/tokens with a single range check. However, my use case seems to be already covered (optimized well), so I'd rather avoid speculation.

Thank you.

[safasofuoglu]

Another attempt that doesnt optimize well:

let _ = inp.parse(any().repeated().at_most(len));

[zesterer]

FWIW the main branch does support parting numbers with lexical (see the docs of the number module) and this will be included in the next alpha release.

There is an internal InputRef::skip_bytes that works for inputs that implement SliceInput<Offset = usize>, used by the regex and lexical-numbers features. However, it performs no validation and hence has the capacity to be unsound if exposed publicly (which reminds me: I need to mark it as unsafe).

I could probably implement a Input::skip_n method that's exposed via InputRef with a default unoptimised implementation that just called next_maybe repeatedly though. Would this suit your needs?