Red Hat Bugzilla – Bug 1313974
Additional test failures
Last modified: 2017-08-08 05:58:01 EDT
Description of problem:
In RHSCL rh-ror42 are additional failing tests than in Fedora. Test results are also very unstable.
Version-Release number of selected component (if applicable):
Make a scratch-build of rh-ror42-rubygem-railties with deleted "grep" of the test execution.
Steps to Reproduce:
1. rhpkg co rubygem-railties && cd rubygem-railties && \
git checkout rhscl-2.2-rh-ror42-rhel-7
2. delete " | grep .*" in %check
3. rhpkg scratch-build
Approx. "1111 runs, 2331 assertions, 210 failures, 396 errors, 0"
Ideally "1012 runs, 2433 assertions, 129 failures, 330 errors, 0"
Test suite is very unstable, test results may vary in every run.
Have you identified a fix for this?
Sorry, no. I have found no fix.
However, I have found out that in Fedora the tests are also very unstable and are disabled since 2015-02-12.
The tests are run (in Fedora/RHSCL) in a fake framework root, which is trying to work around the Bundler usage for the tests.
Additionally, after updating to 4.2.6, the tests do not run at all, due to invalid dependency path(s).