The google codelab Android Room with a View - Kotlin has the following snippet:
class WordViewModel(application: Application) : AndroidViewModel(application) {
// ...
private val coroutineContext: CoroutineContext
get() = parentJob + Dispatchers.Main
private val scope = CoroutineScope(coroutineContext)
// ...
}
And from what I understand from this answer, the custom getter is evaluated everytime, whereas the assignment is evaluated at the time of constructing only. So in effect, scope would take a value that wouldn't change later, so what use is the custom getter for coroutineContext?
I think in this example we can get rid of
private val coroutineContext: CoroutineContext
get() = parentJob + Dispatchers.Main
and just write
private val scope = CoroutineScope(parentJob + Dispatchers.Main)
So the result code will look like:
class WordViewModel(application: Application) : AndroidViewModel(application) {
private var parentJob = Job()
private val scope = CoroutineScope(parentJob + Dispatchers.Main)
// ...
}
I guess it is a matter of style to write getter in this case. Nothing will change if we remove it.
The definition of coroutineContext as a computed property (or custom getter) makes more sense when you are defining a CoroutineScope in a component with lifecycle (i.e. Android Activity). The example in the Javadoc is self-explanatory:
class MyActivity : AppCompatActivity(), CoroutineScope {
lateinit var job: Job
override val coroutineContext: CoroutineContext
get() = Dispatchers.Main + job
override fun onCreate(savedInstanceState: Bundle?) {
super.onCreate(savedInstanceState)
job = Job()
}
override fun onDestroy() {
super.onDestroy()
job.cancel() // Cancel job on activity destroy. After destroy all children jobs will be cancelled automatically
}
}
In this case, you are creating the Job in a lifecycle method, that's why you need the computed property to return the coroutineContext with the Job instance created in onCreate.