Accumulative timeout in ToTry

[dfakhritdinov]

Some cases like skafka #125 required applying ToTry multiple times while accumulating total execution time and firing IO.timeout on its exhaustion.

That can be achieved by implementing factory ToTryOf with API like:

import cats.effect.IO
import cats.effect.unsafe.IORuntime

sealed trait ToTryOf[F[_]] {
  def disposable(timeout: FiniteDuration): F[ToTry[F]]
}

object ToTryOf {
  def ioToTryOf(implicit runtime: IORuntime): ToTryOf[IO] = ???
}

[t3hnar]

that's definitely not a generic thing, we might have something like that in skafka directly

However I still do not understand how you can use that?

For instance in skafka we combine all things smart enough so we do block as rare as possible.

[dfakhritdinov]

The idea behind my proposal is in creating ToTry that will accumulate time taken to execute all effects and raise error if their total execution time exceeds timeout, example:

val toTryOf: ToTryOf[IO] = ???

val foo: Try[Unit] = for {
  implicit0(toTry: ToTry[IO]) <- toTryOf. disposable(10.seconds)
  _ <- IO.sleep(4.seconds).toTry // success
  _ <- IO.sleep(4.seconds).toTry // success
  _ <- IO.sleep(4.seconds).toTry // timeout exception as total execution time is ~12 seconds
  // all further usages of the `toTry` will raise timeout exception immediately 
} yield {}

Now, after taking more detailed look at skafka' RebalanceListener1 I got doubts that the proposal makes sense -- Kafka consumer, available as part of rebalance listener provides access to lots of blocking operations that includes I/O. Time taken by the consumer operations must also be considered in calculation of timeout, that's out of ToTry scope.