
The abstract base docs publication job already defines the post.yaml then publish.yaml playbooks for these jobs. That order is important because we need post.yaml to fetch the sphinx output before we run publish.yaml. However, when we set post-run to publish.yaml in the child jobs that means publish.yaml runs first and fails. Then we continue to run the parent playbooks in the right order (post.yaml then publish.yaml) and successfully update docs. But at this point we've already recorded the job as a failure. Fix this erroneous failure mode by removing the post-run publish.yaml playbook from the child jobs. We can just let the parent defs run and that should ensure the correct order. Change-Id: I0102c0fa2af437cedc129e2fe587ac4314d593e3
Description
Languages
Jinja
51.5%
Python
48.5%