What’s the difference between creating a property with a getter function vs one without. Are these 2...

# android

[voben]

What’s the difference between creating a property with a getter function vs one without. Are these 2 actually the same thing?

val myLiveData: LiveData<String> = _backingLiveData

val myLiveData2: LiveData<String> 
      get() = _backingLiveData

[Casey Brooks]

The first one caches the value in a backing field, and is executed once at object creation. The second one does not have a backing field, is lazily evaluated, and gets executed on each access

[Casey Brooks]

Here’s a playground demonstrating that behavior https://pl.kotl.in/hJgjq9XSh

[John]

The second one would be the same as

fun getMyLiveData(): LiveData<String> { 
   return _backingLiveData
}

[arekolek]

The first one caches the value in a backing field, and is executed once at object creation. The second one does not have a backing field, is lazily evaluated, and gets executed on each access

That’s not a good way of looking at it. It makes it sound like the first one is better, while it’s not. See https://kotlinlang.org/docs/reference/properties.html#backing-properties


Here’s the decompiled bytecode for that example:

private final MutableLiveData<String> _backingLiveData = new MutableLiveData<>();

private final LiveData<String> myLiveData = _backingLiveData;

public final LiveData<String> getMyLiveData() {
  return myLiveData;
}

public final LiveData<String> getMyLiveData2() {
  return _backingLiveData;
}

You can see that myLiveData uses more memory, but doesn’t have any performance benefit.

[nrobi]

If you use proguard/R8, the first option may be still optimized