fix: change storage to use sync queue, fix retry issues

[liuyang1520]

Summary

• fix: change storage to use sync queue
• fix: retry issues

When multiple Task are created, the order of execution doesn't seem to be guaranteed: https://forums.swift.org/t/task-is-order-of-task-execution-deterministic/51553/30, namely FIFO.

After some investigation, I think @globalActor might be a potential solution. However, as the documentation is limited, and I don't have a good example to verify it. I change the storage to use dispatch queue (old generation API). Example I tried with @globalActor:

...
    @SomeGlobalActor
    func test_mainActor_taskOrdering() async {
        var counter = 0
        var tasks = [Task<Void, Never>]()
        for iteration in 1...1000 {
            tasks.append(Task {
                print(Thread.current)
                counter += 1
                XCTAssertEqual(counter, iteration) // often fails
            })
        }
        
        for task in tasks {
            _ = await task.value
        }
    }
...

@globalActor
public struct SomeGlobalActor {
  public actor MyActor { }

  public static let shared = MyActor()
}

Also test the error cases for handlers, found minor issues and fix here.

Checklist

• Does your PR title have the correct title format?
• Does your PR have a breaking change?:

[github-actions]

🎉 This PR is included in version 1.0.0 🎉

The release is available on:

• GitHub release
• v1.0.0

Your semantic-release bot 📦🚀