Running processes and process groups

There are three methods for running a process:

• The run method is the simplest one, it starts the process and then blocks the current fiber until it terminates
• The start method starts the process and returns a fiber packed into a resource. The fiber finishes when the process terminates. Canceling the fiber terminates the process.
• The startProcess method returns a RunningProcess[O, E] interface that allows advanced some operations

Similarly for a process group, there is a run, a start and a startProcessGroup method but with different result types.

Let’s see some examples!

implicit val runner: ProcessRunner[JVMProcessInfo] = new JVMProcessRunner 

val process = Process("echo", List("hello"))

val result1 = process.run()
val result2 = process.start().flatMap { fiber =>
  fiber.join
}

val result3 = 
  for { 
    runningProcess <- process.startProcess()
    _ <- runningProcess.kill()
  } yield ()

Both RunningProcess and RunningProcessGroup has the following methods:

• waitForExit() waits until the process terminates
• terminate() sends SIGTERM to the process
• kill() sends SIGKILL to the process

In addition RunningProcess also defines an isAlive check.

Process execution result

The result of a process is represented by ProcessResult[O, E] defined as follows:

trait ProcessResult[+O, +E] {
  val exitCode: ExitCode
  val output: O
  val error: E
}

The type and value of output and error depends on what redirection was defined on the process.

Process group execution result

The result of a process group is represented by ProcessGroupResult[O, E]:

trait ProcessGroupResult[+O, +E] {
  val exitCodes: Map[Process[Unit, Unit], ExitCode]
  val output: O
  val errors: Map[Process[Unit, Unit], E]
}

The keys of the maps are the original process values used in the piping operations.