Is this any better than having using the backing property that we usually see? This is way has one l...

# android-architecture

[Nick]

Is this any better than having using the backing property that we usually see? This is way has one less property to instantiate so would that make this more memory efficent?

protected val titleLiveData = MutableLiveData<String>()
    
    fun getTitleLiveData(): LiveData<String> {
          return titleLivaData
    }

[Nick]

this._titleLiveData = new MutableLiveData();
this.titleLiveData = (LiveData)this._titleLiveData;

This is what it compiles down to if there is a backing property, since it’s just copying it over its just as efficent, right?

[Nick]

compiling down to bytecode, it shows that the traditional approach with the backing field is more efficent.
Traditional approach with backing field:

LINENUMBER 33 L0
    ALOAD 0
    GETFIELD com.blahblahblah.MyViewModell.title : Landroidx/lifecycle/LiveData;
    ARETURN

Approach with the getter function:

LINENUMBER 29 L0
    ALOAD 0
    GETFIELD com.blahblahblah.MyViewModell._title : Landroidx/lifecycle/MutableLiveData;
    CHECKCAST androidx/lifecycle/LiveData
    ARETURN

The traditional approach doesn’t have the extra cast in it. Memory is the same it appears.