How can I make sure to start collecting to a `Muta...

# coroutines

[mbonnin]

How can I make sure to start collecting to a MutableSharedFlow before emitting something ?

val mutableFlow = MutableSharedFlow<Int>()
      val sharedFlow = mutableFlow.asSharedFlow()

      // start listening to the sharedFlow
      launch {
        sharedFlow
            .collect {
              // by the time we start collecting, the first item is gone already
              // and nothing is received
              println("$it")
            }
      }
      // emit something 
      mutableFlow.emit(1)

Adding a delay before emitting works but doesn't sound great and I don't see a way to enforce a before-after relationship there

[Tijl]

a simple yield() instead of delay might do the trick, but you can also collect the .subscriptionCount on your mutableFlow until it is 1

[mbonnin]

collect the .subscriptionCount

This will ultimately be called from different threads so that sounds dangerous

[Tijl]

if you’ve collected 1 there, you can be sure someone started collecting, regardless of threads

[Tijl]

if you want to be sure they also did not stop, that’s quite a bit broader of a requirement

[mbonnin]

It's a broadcast scenario where a collector will filter only a subset of all emitted items

[mbonnin]

Each collector also "controls" when the item start being emitted

[mbonnin]

But I want to make sure they don't miss any

[Tijl]

https://kotlin.github.io/kotlinx.coroutines/kotlinx-coroutines-core/kotlinx.coroutines.flow/on-subscription.html if you do this one per collect you can see when they start

[jw]

launch(start = UNDISPATCHED) { .. }

[jw]

the coroutine will eagerly start in the current stackframe until it hits a suspension point (in this case, the collect call)

[babel]

passing replay = 1 into MutableSharedFlow would also get you the println

[mbonnin]

@babel yea but replay = 1 doesn't work anymore if there can be n emitted items as I don't know 'n' beforehand. Overall I like the onSubscription() solution. Looks like it was made especially for this use case