I am trying to access the C-style vector from Rust (for Rizin's Rust bindings), currently the conversion happens like this:
/// A generic converter from RzVector to Vec (with a type conversion function)
macro_rules! rzpvector_to_vec {
($fn_name:ident, $base_type:ident, $return_type:ident, $conversion_fn:ident) => {
fn $fn_name(core: &RizinCore, pvec: *mut RzPVector) -> Vec<$return_type> {
let mut vec: Vec<$return_type> = Vec::new();
unsafe {
while !rz_pvector_empty(pvec) {
let element = rz_pvector_pop_front(pvec) as *mut $base_type;
if let Some(el) = $conversion_fn(core, element) {
vec.push(el);
}
}
}
vec
}
};
}
Where RzPVector is the C (see complete implementation of it in librz/util/vector.c and librz/include/rz_vector.h) structure like this:
typedef struct rz_vector_t {
void *a; //< a continuous chunk of memory where all elements are stored sequentially
size_t len;
size_t capacity;
size_t elem_size;
RzVectorFree free;
void *free_user;
} RzVector;
// RzPVector directly wraps RzVector for type safety
typedef struct rz_pvector_t {
RzVector v;
} RzPVector;
Is there any way to somehow impose Rust Vec-like (at least simple iteration, etc) traits on top of this structure without copying? Because creating a new Vec works fine for small structures but copying gets problematic on really huge RzPVectors.