Would `heapless` accept a PR for a `Deque`/`RingBuffer` that operates on a mutable slice instead of allocating a buffer of a fixed size?

[ColinFinck]

I basically need smoltcp's RingBuffer in a dedicated library, because I want to use it in a non-networking context. You currently cannot use it without adding a dependency to smoltcp and its network features.

One idea is to move it from smoltcp to a dedicated crate, and I have sent out an inquiry whether they would accept a PR in that direction.
Another idea is to use a similar data structure from another crate. I have looked at the Deque of heapless, but it requires a capacity type parameter. smoltcp's RingBuffer asks the user to provide a mutable slice. The same RingBuffer can therefore operate on buffers of arbitrary size. This is exactly what I need for my application.

Now I would be happy with any data structure that may be used like a ring buffer, as long as it operates on mutable slices. Be it another Deque-like struct that shares much of its API (what would be a good name for that?), or taking over the original "RingBuffer" with its API.

Before I come up with a PR: Which of these options would the heapless crate accept?

[zeenix]

I have sent out an inquiry whether they would accept a PR in that direction.

Could you please point out to where you asked? I don't see anything here. 🤔

have looked at the Deque of heapless, but it requires a capacity type parameter.

All of the heapless API is geared toward fixed-sized buffers on the stack. So even if we were to add an API to allow user to provide a mutable slice, we would still need to support the existing behaviour/API and therefore still have the const generic on types.