Comment 30 for bug 1979244

You misunderstood the test, the test has phased-security 2 in security and phased-security 3 is phasing in updates, at 0 to yield an easy test that it's really "not for us".

We don't have special casing for phased-update-percentage set on security updates. The behavior here is identical to update-manager and while I agree that it might make sense to handle that I feel like we can discuss that later rather than block this critical SRU now.

Surely everyone is aware that update-manager does not respect phasing on security updates already and takes care.

This SRU really is just "do what update-manager does, it worked so far". I prefer the pinning based implementation long term as it allows switching candidates (so it installs a good version from security rather than the bad version from updates) and works for install as well but needs more work in both launchpad dominator to keep previous versions in updates and apt to propagate candidate selections across a source package (set).