SELinux RPM scriplet issue annoucement

[Michael Schwendt]

On Sun, 19 Jan 2014 19:15:35 +0100, drago01 wrote:

> So it happened .. how do we prevent it in the future? How did it pass testing?

A first +1 vote 22 hours _before_ it entered the updates-testing repo.
A second +1 vote eight hours _before_ it entered the updates-testing repo.
A third +1 vote and automatic push to stable two hours after it had entered
updates-testing:

  https://admin.fedoraproject.org/updates/FEDORA-2014-0806/selinux-policy-3.12.1-116.fc20

It has not been offered long enough for more testers to give it a try.
For example, here it had not been picked up by the nearby mirror yet when
the karma threshold was reached. By the time it had arrived and the first
scriptlet errors were noticed and required a closer look to find the culprit,
the update had been pushed to stable already.

How to prevent it from happening in the future? The update criteria for
the so-called critical path packages could be made more strict. A minimum
time for updates to stay in the updates-testing repo. A higher karma
threshold probably won't be sufficient, if testers aim at speed instead
of safety.