Hello all, I am having a difficult time wrapping my head around a concurrency issue when using coroutines. As a small example I have data coming in from the system that I save locally to a database. At some frequency I want to push those results in a batch request to a server. I would like to ensure that only one push “task” is running at time to reduce load on the server. Will throw some code in this thread…

coroutineScope { launch { ... } }

is basically equivalent to

...

, the job is always complete by the time the

suspend fun

returns

Ok, that makes sense, thank you. So something like this?

fun onEvent(event: Event) {
   // Called on Main thread
   lifecycleScope.launch {
      eventRepository.save()  // Executes on IO thread
      pushJob?.join()
      pushJob = launch {
         Log.v("Test", "start push")
          ...
          locationRepository.pushLocations()  // Executes on IO thread
          ...
          Log.v("Test", "end push")
      }
   }
}

that could work but why not just use a

Mutex

?

I would like to allow the save to be non blocking, but yeah I think this would achieve what I am after as the save is inside the coroutine and new calls to onEvent would just launch a new coroutine, save and then block if necessary

fun onEvent(event: Event) {
   // Called on Main thread
   lifecycleScope.launch {
      eventRepository.save()  // Executes on IO thread
      mutex.withLock {
         Log.v("Test", "start push")
          ...
          locationRepository.pushLocations()  // Executes on IO thread
          ...
          Log.v("Test", "end push")
      }
   }
}