Code Quiz

Resilient Fetch With Dedup And Backoff

A request-coordinating actor combines retries, exponential backoff, and in-flight deduplication — but its backoff never actually delays.

Codeswift

actor RequestCoordinator {
    private var inFlight: [URL: Task<Data, Error>] = [:]

    func data(for url: URL, maxAttempts: Int = 3) async throws -> Data {
        if let task = inFlight[url] {
            return try await task.value            // dedup: join in-flight request
        }
        let task = Task { try await self.fetchWithRetry(url, maxAttempts: maxAttempts) }
        inFlight[url] = task
        defer { inFlight[url] = nil }
        return try await task.value
    }

    private func fetchWithRetry(_ url: URL, maxAttempts: Int) async throws -> Data {
        for attempt in 0..<maxAttempts {
            do {
                let (data, resp) = try await URLSession.shared.data(from: url)
                guard let http = resp as? HTTPURLResponse,
                      200..<300 ~= http.statusCode else {
                    throw URLError(.badServerResponse)
                }
                return data
            } catch {
                try Task.checkCancellation()
                let delay = pow(2.0, Double(attempt))          // seconds: 1, 2, 4
                try await Task.sleep(nanoseconds: UInt64(delay))
            }
        }
        throw URLError(.cannotLoadFromNetwork)
    }
}

The dedup and cancellation logic are correct, but the backoff is broken. What is the bug and its fix?

Next →Retry Wrappers vs. Resource Timeouts