Imagine we have a simple helper class which only consists of public static methods.
1 2 3 | public class Helper { public static void doSomething() {}} |
And the test:
1 2 3 4 5 6 7 8 | import org.junit.Test;public class HelperTest { @Test public void testDoSomething() { Helper.doSomething(); }} |
Our code coverage tool won't show us a 100% coverage, because the (default) constructor is not invoked. For example Intellij IDEA only shows 50%.
One solution would be to simply invoke the default constructor with new Helper.doSomething(). But we don't want to have objects of this class and therefore override the default constructor with a private one:
1 2 3 4 | public class Helper { private Helper() {} public static void doSomething() {}} |
The problem is, that now even the test cannot instantiate the class. And as we are testdriven, the question is: Should the decision to create a class without any objects be motivated by tests, too? In my opinion yes. I don't want some other programmer to remove the private constructor. At least a test shall break.
We reach this by the following test:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 | import static org.junit.Assert.assertTrue;import java.lang.reflect.Constructor;import java.lang.reflect.Modifier;import org.junit.Test;public class HelperTest { @Test public void testDoSomething() throws Exception { Helper.doSomething(); } @Test public void testPrivateConstructor() throws Exception { Constructor constructor = Helper.class.getDeclaredConstructor(); assertTrue("Constructor is not private", Modifier.isPrivate(constructor.getModifiers())); constructor.setAccessible(true); constructor.newInstance(); }} |
Now we have a guard to the design decision of a private constructor and our tests are back on 100% coverage again.