I'm using Tuple2 in my Java code, and I was wondering if there was a difference between accessing the values via the getter or just getting the variables directly.
Tuple2<String,String> tuple = new Tuple2<>("Hello", "World");
//getting values directly
String direct = tuple._1;
String indirect = tuple._1();
The first loads a field where the second invokes the method using
invokeVirtual opcodes relatively. The generated byte-code looks like
13: getfield #6 // Field scala/Tuple2._1:Ljava/lang/Object; 16: checkcast #7 // class java/lang/String 19: astore_2 20: aload_1 21: invokevirtual #8 // Method scala/Tuple2._1:()Ljava/lang/Object; 24: checkcast #7 // class java/lang/String
And the difference is the difference between a field read and a method invocation, ie the
JIT compiler is happy to inline the method and it wouldn't matter much performance wise.