Copy vs Move Semantics

When a value moves around your program, Mojo can either copy it or move it. Knowing which happens helps you reason about cost. 🔁

A copy makes a brand-new, independent value. Both the original and the copy live on, each owning its own data.

A move relocates a value instead of duplicating it. The data is handed over, so the source is no longer used afterward.

A copy may duplicate heap data, but a move just transfers ownership of what is already there. For big values that saves real work.

Plain assignment in Mojo defaults to a copy, so each variable owns its own value. This keeps value semantics safe and predictable.

You request a move with the transfer operator, the caret sigil. It tells Mojo to hand the value over rather than duplicate it.

After a move, the source variable is consumed. Mojo will not let you read it again, which prevents using stale or invalid data.

By contrast, after a copy both variables are fully usable. Each one is independent, so editing one never touches the other.

Mojo can also turn a copy into a move automatically when the source is no longer needed, an optimization that quietly saves work.

Thinking about copy versus move lets you avoid duplicating large buffers by accident. It is the first step toward efficient Mojo code.

Mojo gives you safe defaults but lets you opt into a move when you want speed. You choose the behavior that fits each moment.

Lets test copy versus move.

A copy makes an independent duplicate, while a move transfers ownership cheaply and spends the source. Now you can predict both. 🎯