#9169 Block unretirement commits
Closed: Fixed by mohanboddu. Opened by churchyard.

  • Describe the issue

See https://pagure.io/releng/issue/9168#comment-621173

We should really prevent maintainers to push retirement-revert commits. For two reasons:

  • maintainers often get a false feeling that they can unretire the package by reverting (on accidental retirements etc.) -- that is false, releng needs to unblock the package in Koji
  • releng wants to add a ticket link to the commit message for bookkeeping

How about when pdc is updated (by seeing a dead.package commit come accross the bus) we have it update pdc that the package is EOL on the previous day (ie, it's already eol)).

Then the hook we have to prevent pushing to eol in pdc things would stop people from reverting the retirement?

We would still need to make sure releng can push to unretire things, but that might already be the case.

Yes, that works for the case of unretirement commits that happen after pdc update.

It does'n however solve the problem with immediate reverts. That is tracked in https://pagure.io/releng/issue/8851

Metadata Update from @mohanboddu:
- Issue tagged with: meeting

It does'n however solve the problem with immediate reverts. That is tracked in https://pagure.io/releng/issue/8851

It does also solve it, since people cannot revert/push the unretirement commit immediately.

I know its not the best solution, but its the easiest solution we got that will work.

It does also solve it, since people cannot revert/push the unretirement commit immediately.

They can, if they push it sooner than the automation triggers the change in PDC.

It does also solve it, since people cannot revert/push the unretirement commit immediately.

They can, if they push it sooner than the automation triggers the change in PDC.

I know its not super optimal, but it will be a gap fix until we come up with a good solution.

I hotfixed it using https://pagure.io/releng/issue/9169#comment-625104

Closing the ticket as there is no better way to fix it.

Metadata Update from @mohanboddu:
- Issue close_status updated to: Fixed
- Issue status updated to: Closed (was: Open)

Thanks.

Metadata