Ah, I have answered on Codeberg for once :) Am Fri, Mar 20, 2026 at 09:25:02AM -0400 schrieb Jason Conroy: > I think Steve has a good point here. When talking about GCDs for example, > "consensus" has a rather specific definition[1] where a single disapproval > prevents a proposal from advancing. Andreas, are you suggesting the same for > package removals, or a weaker form of consensus like "a generally accepted > decision"[2]? Also, are the consensus-reachers limited to project members > (like with GCDs) or is anyone with a Codeberg account welcome to vote? I do > think that a single sentence in the GCD clarifying the intent would help.
I was proposing that a single disapproval prevents a deprecation from taking place. But I realised that this makes no sense for broken packages, where the only way of going forward is to repair it. So I have pushed a change (called "v3") where to prevent a broken package from being removed, one should suggest to repair it. And for building leaf packages, a removal requires consensus; in the past, a single user stating "I still play this game" has stopped removals, and I think we should keep this policy. There is no vote, there is no need to be a team member, anyone can voice a significant concern. For packages inside the graph, this does not hold (we do not want to keep these pesky Python 2 games forever ;-)), but there the barrier for removal is very high anyway. Andreas
