Concurrency

In the last posts we looked at running work in parallel with async let and TaskGroup, and then how Sendable ensures the data we share across tasks is safe. The next piece of Swift’s structured concurrency story is actors. Actors provide isolation for mutable state, making them one of the core building blocks for avoiding data races. What is an Actor? An actor is like a class, but with one key difference: only one task at a time can interact with its mutable state. This guarantees thread safety without you having to add locks manually. ...

In the previous post, we looked at running work in parallel with async let and TaskGroup. But once tasks start running in parallel, there’s a bigger question: what happens to the data they share? Swift’s answer is the Sendable protocol. What is Sendable? Sendable is a marker protocol that tells the compiler a type can safely be passed across concurrency domains like tasks and actors. If a type isn’t Sendable, Swift will stop you from moving it into a concurrent context because that might cause a data race. The compiler does a lot of this checking automatically, but sometimes you’ll need to be explicit. ...

In the previous post, we loaded three independent requests sequentially. Even though the operations didn’t depend on each other, each one waited for the previous to finish To run them in parallel while keeping structured concurrency, Swift gives us two main tools: async let and TaskGroup. When to use async let Let’s start with async let. It allows you to fire off multiple operations in parallel within the same task context: ...

One of the biggest changes in Swift since I’ve been away is how we handle concurrency. Back in iOS 15, we got our first taste of Swift Concurrency with async/await. At the time, most of us were still relying heavily on Combine or traditional closures to handle asynchronous work. I had played around with it a little back then, but its practical use was limited because it couldn’t be deployed widely on older iOS versions. ...