How is k commonly selected for KNN?

A. By setting k equal to the number of features automatically B. By comparing validation or cross-validation performance C. By using the test set repeatedly D. By choosing the largest possible k