How to run a test job when some build jobs have failed?
Hi,
we are trying to write a .gitlab-ci.yml file which supports two platforms (linux and windows) and has two stages (build and test). Following common advice, we used platform-specific runner tags to assign jobs to a corresponding platform runner. The problem is, if a build stage job fails on one platform (but the other one succeeds), it automatically blocks all test stage jobs.
Is there currently a way or a proposed feature to enable execution of the test job for the platform on which a build job has succeeded, even if build jobs for other platforms have failed (based on tags or otherwise)?
Of course, I could put test script commands into my build jobs, but that would defeat the whole idea of having stages.
We are running a local instance of Gitlab Community Edition 8.15, but can upgrade.