The only method I can find is into_usizes which creates a new allocation.

I can manually accomplish this via iterating over all indices and using .filter(|i| matches!(bitset.get(i), Some(true)) but it seems like this is potentially not the most efficient implementation.

Would it be possible to add a method for this? Perhaps one that returns the iterator currently used internally in into_usizes?

        v.storage
            .into_iter()
            .flat_map(|x| x.to_array())
            .flat_map(|x| { (0..A::Item::ELEMENT_BIT_WIDTH).map(move |i| (x.wrapping_shr(i as u32)) & A::Item::ONE_ELEMENT != A::Item::ZERO_ELEMENT) })
            .take(v.nbits)
            .enumerate()
            .filter(|(_, b)| *b)
            .map(|(i, _)| i)