BitBox is a heap-allocated BitSlice region. It is a distinct type because the implementation of bit-slice pointers means that Box<BitSlice> cannot exist. It can be created by cloning a bit-slice into the heap, or by freezing the allocation of a BitVec
As with BitSlice, this takes a pair of BitOrder and BitStore type parameters to govern the buffer’s memory representation. Because BitSlice is unsized, BitBox has almost none of the Box API, and is difficult to use directly.
BitBox, like &BitSlice, is an opaque pointer to a bit-addressed slice region. Unlike &BitSlice, it uses the allocator to guarantee that it is the sole accessor to the referent buffer, and is able to use that uniqueness guarantee to specialize some BitSlice behavior to be faster or more efficient.
BitBox is, essentially, a NonNull<BitSlice<T, O>> pointer. The internal value is opaque and cannot be inspected or modified by user code.
If you attempt to do so, your program becomes inconsistent. You will likely break the allocator’s internal state and cause a crash. No guarantees of crash or recovery are provided. Do not inspect or modify the BitBox handle value.
The simplest way to construct a BitBox is by using the bitbox! macro. You can also explicitly clone a BitSlice with BitBox::from_bitslice, or freeze a BitVec with BitVec::into_boxed_bitslice.
use bitvec::prelude::*; let a = BitBox::from_bitslice(bits![1, 0, 1, 1, 0]); let b = bitbox![0, 1, 0, 0, 1]; let b_raw: *mut BitSlice = BitBox::into_raw(b); let b_reformed = unsafe { BitBox::from_raw(b_raw) };