Disable test_PlaceholderOnTopStackSwitch() on MacOS#3944
Disable test_PlaceholderOnTopStackSwitch() on MacOS#3944trancexpress wants to merge 1 commit intoeclipse-platform:masterfrom
Conversation
HeikoKlare
left a comment
There was a problem hiding this comment.
Disabling the test means that the functionality is potentially broken on macOS or at least will further degrade/break due to lack of test coverage. In case that's fine for those consumers using macOS, I am fine with disabling the test. It might be reasonable to create an issue (or rephrease the existing one for the failing test) to document the disabled tests and the potentially broken functionality.
| @Test | ||
| public void test_PlaceholderOnTopStackSwitch() { | ||
| // test fails stably on MacOS I-builds, its unclear how to fix it, see: https://github.com/eclipse-platform/eclipse.platform.ui/issues/3893 | ||
| Assumptions.assumeFalse(Platform.OS.isMac()); |
There was a problem hiding this comment.
If I am not mistaken, with JUnit 5 it should be possible to use @DisabledOnOs(OS.MACOS).
I'm mostly worried because the test seems to work with As far as I-builds are concerned, we already don't know if the respective functionality works or not - the test fails in each build. But technically we would know if a PR breaks the test. I don't know how to disable the test for I-builds only though... Maybe Maybe we can add something in Here are the ENV variables set during the I-build: https://download.eclipse.org/eclipse/downloads/drops4/I20260428-1800/testresults/consolelogs/ep440I-unit-macosx-aarch64-java21_macosx.cocoa.aarch64_21_consolelog.txt No idea where to look for this during the CI job: https://github.com/eclipse-platform/eclipse.platform.ui/actions/runs/25103050435/job/73557140458?pr=3944 |
61cbeaf to
d190a4c
Compare
|
See test jobs from the PR, which should fail with the value of
All of them fail without a message, i.e. This is more complicated than just disabling the test. But if it passes in PR jobs, then better its disabled only for I-builds... Best of course would be if the test gets fixed, but I have no MacOS machine and can't do much about that. |
|
We already have (SWT) tests that check whether they are executed in Jenkins. Maybe you could adopt that: https://github.com/eclipse-platform/eclipse.platform.swt/blob/b40fe9428f8e3f280e7b6fa296c8bd809138ade7/tests/org.eclipse.swt.tests/JUnit%20Tests/org/eclipse/swt/tests/junit/SwtTestUtil.java#L135-L137 |
I can check if its set to some other value during the PR jobs, but I'm not sure how reliable that will be? Though I'm also not sure how reliable checking |
|
The SWT tests just check if |
|
It does seem that one of the jobs has I'm not sure which job that is though, since the Windows/Linux/MacOS jobs show a pass status... From my POV we can still ignore if |
Sounds reasonable. So guarding the test with |
PartOnTopManagerTest.test_PlaceholderOnTopStackSwitch() fails stabily in MacOS I-builds. Since its unclear what is wrong, the test is now disabled to reduce noise in test results. See: eclipse-platform#3893
| * {@link #JENKINS_JOB_NAME_ENV_VAR} is set. Jenkins jobs are: | ||
| * <ul> | ||
| * <li>I-builds</li> | ||
| * <li>GitHub PR test job, excluding the Windows/Linux/MacOS test jobs ran with |
There was a problem hiding this comment.
I'm not sure how to word this...
|
@iloveeclipse please review this. The respective test has been failing in every I-build for about 3 weeks now, for MacOS ARM. |
PartOnTopManagerTest.test_PlaceholderOnTopStackSwitch()fails stabily in MacOS I-builds.Since its unclear what is wrong, the test is now disabled to reduce noise in test results.
See: #3893