In this example https://github.com/Kotlin/kotlinx.coroutines/blob/master/kotlinx-coroutines-core/jvm/test/guide/example-exceptions-03.kt why do we need to call “yield” at line 19 and 23?

runBlocking

sets up a single threaded dispatcher running inside

runBlocking

on the callers thread.

val child = launch {

sets up the child coroutine but the child coroutine doesn't actually run.

yield

gives up the thread so that child coroutine can actually start. Without the first

yield

, the child coroutine wouldn't ever have a chance to run before

child.cancel()

was called. "Child is cancelled" would never be printed because lambda would never run at all due to it being cancelled before it even started. The second

yield

appears to be pointless since

join

gives up the thread until the child coroutine has finished.