Remarks on some existence theorems for optimal control