I have a Repository defined as the following.
class StoryRepository {
private val firestore = Firebase.firestore
suspend fun fetchStories(): QuerySnapshot? {
return try {
firestore
.collection("stories")
.get()
.await()
} catch(e: Exception) {
Log.e("StoryRepository", "Error in fetching Firestore stories: $e")
null
}
}
}
I also have a ViewModel like this.
class HomeViewModel(
application: Application
) : AndroidViewModel(application) {
private var viewModelJob = Job()
private val uiScope = CoroutineScope(Dispatchers.Main + viewModelJob)
private val storyRepository = StoryRepository()
private var _stories = MutableLiveData<List<Story>>()
val stories: LiveData<List<Story>>
get() = _stories
init {
uiScope.launch {
getStories()
}
uiScope.launch {
getMetadata()
}
}
private suspend fun getStories() {
withContext(Dispatchers.IO) {
val snapshots = storyRepository.fetchStories()
// Is this correct?
if (snapshots == null) {
cancel(CancellationException("Task is null; local DB not refreshed"))
return@withContext
}
val networkStories = snapshots.toObjects(NetworkStory::class.java)
val stories = NetworkStoryContainer(networkStories).asDomainModel()
_stories.postValue(stories)
}
}
suspend fun getMetadata() {
// Does some other fetching
}
override fun onCleared() {
super.onCleared()
viewModelJob.cancel()
}
}
As you can see, sometimes, StoryRepository().fetchStories()
may fail and return null
. If the return value is null
, I would like to not continue what follows after the checking for snapshots
being null
block. Therefore, I would like to cancel that particular coroutine (the one that runs getStories()
without cancelling the other coroutine (the one that runs getMetadata()
). How do I achieve this and is return
-ing from withContext
a bad-practice?