We had a discussion at work about where to put the try keyword when using async let.
Let’s say we have a function defined like this:
func bar() async throws -> Response?
A colleague wanted to use async let to start the call, then await the response later in their task. They also wanted to utilise the try? operator to just let the expected return value become nil if the call throws an error. They wrote their task like this:
Task {
async let foo = try? bar()
// ... other stuff
self.response = await foo
}
This didn’t sit right with me, because in my experience you would always put the try with the await. I am wondering why Xcode didn’t throw an error here, so I experimented a bit and found that if we use a try without the ? in the same way, we DO get an error:
Task {
async let foo = try bar()
// ... other stuff
self.response = await foo // This line errors, see below
}
That is the error I expected for the above case, and to me that signifies that the try on the async let declaration does nothing. However, that still leaves me wondering why my colleague didn’t get a compiler error with their code.
Is this a compiler issue? Or does the try? operator lead to the correct result when an error is thrown? According to my colleague, their code also delivered the expected results, but I don’t know how extensively it was tested.