Reset WebDriver and vncRecordingContainer when Selenium container stops

[tobiasstadler]

Some of our tests look similar to:

void BaseTestClass {

    @Container
    private static final BrowserWebDriverContainer<?> container = new BrowserWebDriverContainer<>("...");

    ...
}

void MyTest1 extends BaseTestClass {

    @Test
    void test1() {
        ...
    }

    ...
}

void MyTest2 extends BaseTestClass {

    @Test
    void test2() {
        ...
    }

    ...
}

The selenium container is started before any test of one class are run and stopped after all test of one class are run. Currently driver is created only once in BrowserWebDriverContainer.java and when the container stops driver.quit() is called. So the second and any subsequent test class will see a "stopped" driver which will fail the tests of that class.

Right now we recreate/start/stop the selenium container manually in @BeforeAll/@AfterAll block, but it would be nice if we could reuse the container definition.

[tobiasstadler]

@kiview Would you please have a look?

[kiview]

Hey @tobiasstadler, sure thing, but can you please update the description of the PR, so that we have the context? Like, which use cases does this unblock?

[tobiasstadler]

@kiview I updated the description

[tobiasstadler]

ping

[kiview]

I am not sure I understand your example correctly. In case of using a JUnit-Rule or the Jupiter-Extension on a per-test-method basis, each test would instantiate a new BrowserWebDriverContainer object.

[tobiasstadler]

@kiview You are right, I described the wrong case. Sorry! I changed the description to actually describe or problem.

[tobiasstadler]

ping

[tobiasstadler]

ping

[bsideup]

Hey @tobiasstadler! Thanks for spotting and fixing it. Do you think you can provide a simple test that will verify that this is no longer an issue? (so that we won't introduce a regression in the future)

[kiview]

Thanks a lot for the contribution @tobiasstadler, we will take care of the test ourselves 🙂

[tobiasstadler]

Thank You!