Handle ControlThrowable exception in 'race' function

[rcardin]

I added the handling of a ControThrowable to the internals of the race function. The ControThrowable exceptions have suppression disabled by default, so it's not possible to accumulate suppressed exceptions.

Feel free to let me know if anything needs clarification or if you notice something needs to be improved.

Closes #213

[adamw]

I'm actually wondering if we shouldn't introduce a third state here. So race (as opposed to raceResult) returns the first success, or returns a failure if all computations fail. So the question is, should we also wait for other computations to fail, if one of the reports throws a ControlThrowable? Maybe you can share a use-case?

I was thinking about the other exceptions that might get thrown. InterruptedException, VirtualMachineError, ThreadDeath, LinkageError are all exceptions which should cause race to finish immediately (without checking if other branches throw exceptions, or not). There's still a bug in the current implementation as to how these are handled, but I'm not 100% sure how to handle thee fatal exceptions in terms of the contract of race.

[rcardin]

Maybe you can share a use-case?
Sure. raise4s uses a ControlThrowable to forward typed errors to the caller, as they were exceptions. So, take the following code for example:

val result: Int raises String = ox.race(
  {
    ox.sleep(500.milliseconds)
    Raise.raise("Something went wrong")
  }, {
    ox.sleep(1.second)
    1
  }
)
Raise.run(result) match {
  case value: Int => println(s"Value: $value")
  case error: String => println(s"Error: $error")
}

I expected the above code to print Value: 1 to the console, waiting for the long task to finish, since the first went in a logic error. I don't know if we should make this behavior universal for ControlThrowables.

but I'm not 100% sure how to handle the fatal exceptions in terms of the contract of race

AFAIK the Fatal exception should not be handled. I agree that if the code throws a fatal exception, the whole race must stop immediately and rethrow the fatal error.

Please let me know your thoughts and how you'd like me to proceed.

[adamw]

Thanks, I've extended the code to also handle other fatal exceptions

[rcardin]

Thanks, Adam. I saw the code you added. It sounds great. I'll keep looking to see if there is any other place in the code with the same problem we spotted here.