observeForever

removeObserver

observeForever

LiveData

MutableLiveData<String>

android:text="@={textLiveData}"

Why would the viewmodel call observeForever on it’s own livedata object?

He asked does calling observeForever on a livedata object leak it after the viewmodel is destroyed

I’m saying that where leaks usually happens is when the observing object has a different lifecycle

removeObserver

Logger.logEvent(text)

dataLogger

twoWayBinding

twoWayBinding

If you use Transformations to react to a two-way binding MutableLiveData, then you need to observe to that Transformations somewhere for it to react to the MutableLiveData, which is why you need to call observeForever in the viewmodel

MutableLiveData

MediatorLiveData

observeForever is reserved for unit testing purposes

@ExtendWith(InstantExecutorExtension::class)
class TestTe {
    @Test
    fun test() {
        val vm = Vm()
        vm.a.observeForever {  }
        vm.a.value = "hi"
    }
}

class Vm {
    val a = MediatorLiveData("")
    private val backingBoolData =  MutableLiveData<Boolean>()
    val boolData: LiveData<Boolean> = backingBoolData
init {

a.addSource(backingBoolData) {
            println("a onChange") // now this is called
            backingBoolData.value = true
        }
    }
}

backingBoolData

class ExampleLiveDataUnitTest {

    @get:Rule
    val rule = InstantTaskExecutorRule()

    val textData = MediatorLiveData<String>()
    val boolData = MutableLiveData<Boolean>()

    var actual = false

    @Before
    fun setUp() {
        textData.addSource(textData) {
            boolData.value = it == "hi"
        }
        textData.addSource(boolData) {
            actual = it
        }
        textData.observeForever {  }
    }

    @Test
    fun addition_isCorrect() {
        val expected = true

        textData.value = "hi"

        assertEquals(expected, actual)
    }
}

observeForever

observeForever

textData.addSource(textData) {
            boolData.value = it == "hi"
        }