ExpectedException fails when assumption fails

[smithkm]

I have tests that Assume.assumeTrue(false) under some circumstances to indicate the test is not relevant. Some of these tests are using an ExpectedException rule to check for other exceptions. Under JUnit 4.11 the test gets ignored as expected, under JUnit 4.12 the rule tries to match the AssumptionViolatedException and fails.

Probably related to the fix for #687

[kcooney]

Thanks for reporting this

If you are using 4.12 I strongly recommend replacing usages of ExpectedException with org.junit.Assert.assertThrows.

[smithkm]

It doesn't appear that's in 4.12. For now I've made it work with some @Ignores but I may try putting in my own assertThrown until 4.13

[kcooney]

Woops, sorry, assertThrows will be added in JUnit 4.13

[kcooney]

Looking at the threads (particularly #720) it sounds like we changed this behavior in 4.11, realized that the new behavior was a bad idea, and reverted to the 4.10 behavior in 4.12.

@smithkm How does your test behave in JUnit 4.10?

[kcooney]

@smithkm It looks like we have a test case for exactly this use case:

https://github.com/junit-team/junit4/blob/master/src/test/java/org/junit/rules/ExpectedExceptionTest.java#L279

This results in an AssumptionFailure (i.e. the test will be effectively ignored)

If you call ExpectedException.expect() before calling assumeTrue() or assumeFalse() then the test will fail (not be ignored).

[allanhaywood]

@kcooney @smithkm Is this still an issue?

[kcooney]

@allanhaywood based on the JUnit self tests, ExpectedException works as expected when there is an assumption failure.

Unless someone provides a minimal, complete and verifiable example demonstrating it, I think we should close it as not reproducible

[marcphilipp]

Optimistically closing. I'd reopen it if someone provides a reproducer.

[allanhaywood]

@kcooney I agree. Thanks for confirming.