Ownership & Moves

How ownership of a Vec moves between variables and into a function, and who frees it

Before anything runs

The compiled program has been loaded into memory, but no function has run yet. Nothing is on the stack and nothing is on the heap

1fn main() {
2    let a = vec![1, 2, 3];
3    let b = a;
4    take(b);
5}
6
7fn take(v: Vec<i32>) {
8    println!("{}", v.len());
9}

How ownership of a Vec moves between variables and into a function, and who frees it

Where this lives in a running program

The idea in one picture

A move is a shallow copy of the stack bytes plus the compiler forbidding the old name, and the optimiser often removes even that copy
Calling .clone() deep-copies the heap data when you really need two independent owners
Rc<T> and Arc<T> allow shared ownership by counting references instead

stdoutempty

$cargorun// waiting for the first println!

Ownership & Moves

How ownership of a Vec moves between variables and into a function, and who frees it

What's happening · 01 / 09

The compiled program has been loaded into memory, but no function has run yet. Nothing is on the stack and nothing is on the heap

Code

1fn main() {
2    let a = vec![1, 2, 3];
3    let b = a;
4    take(b);
5}
6
7fn take(v: Vec<i32>) {
8    println!("{}", v.len());
9}

Ownership & Moves

How ownership of a Vec moves between variables and into a function, and who frees it

Where this lives in a running program

The idea in one picture

A move is a shallow copy of the stack bytes plus the compiler forbidding the old name, and the optimiser often removes even that copy
Calling .clone() deep-copies the heap data when you really need two independent owners
Rc<T> and Arc<T> allow shared ownership by counting references instead