Created attachment 265545 [details] emerge --info I get two test failures: [ FAILED ] 2 tests, listed below: [ FAILED ] MessageLoopTest.FAILS_EnsureTaskDeletion [ FAILED ] MessageLoopTest.FAILS_EnsureTaskDeletion_Chain [ RUN ] MessageLoopTest.FAILS_EnsureTaskDeletion base/message_loop_unittest.cc:440: Failure Value of: a_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:440: Failure Value of: a_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:440: Failure Value of: a_was_deleted Actual: false Expected: true [ FAILED ] MessageLoopTest.FAILS_EnsureTaskDeletion (0 ms) [ RUN ] MessageLoopTest.FAILS_EnsureTaskDeletion_Chain base/message_loop_unittest.cc:455: Failure Value of: a_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:456: Failure Value of: b_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:457: Failure Value of: c_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:455: Failure Value of: a_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:456: Failure Value of: b_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:457: Failure Value of: c_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:455: Failure Value of: a_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:456: Failure Value of: b_was_deleted Actual: false Expected: true base/message_loop_unittest.cc:457: Failure Value of: c_was_deleted Actual: false Expected: true [ FAILED ] MessageLoopTest.FAILS_EnsureTaskDeletion_Chain (1 ms)
The FAILS prefix indicates those are "expected" or "known" failures. The ebuild will die if the tests finish with non-zero exit status. There is also a FLAKY prefix that indicates unreliable test (and we also ignore the result).
The ebuild died on this. If these tests are expected to fail then the ebuild shouldn't die. Did I understand you correctly in thinking that the ebuild should not have died?
(In reply to comment #2) > The ebuild died on this. If these tests are expected to fail then the ebuild > shouldn't die. Did I understand you correctly in thinking that the ebuild > should not have died? Yes, the ebuild should not have died and I think it doesn't for me. Please attach the full build log and reopen.
Created attachment 265561 [details] build.log.bz2
That might be caused by this: * QA: ebuild is exporting $maketype=out/Release/base_unittests * QA: Ebuild should be migrated to use VIRTUALX_COMMAND=out/Release/base_unittests instead. * QA: Setting VIRTUALX_COMMAND to $maketype conveniently for now. I think none of my tests has used the new version of the virtualx eclass. I'll take a more detailed look.
Should be fixed now, thank you for the report.
Fix confirmed. Thanks.