At least 3 PMC member votes are required for release ( https://www.apache.org/foundation/voting.html ) due to foundation policy, but if a committer votes against a release the PMC better be paying attention.
Would argue it should be commit-bit binding for CEPs, though I’d be shocked if there’s ever a case where that particular nuance in vote counting somehow changed the result. > On Jun 13, 2025, at 1:51 PM, Josh McKenzie <jmcken...@apache.org> wrote: > > Doug had a great question that had me digging through gdoc comments from 5 > years ago to try and figure out the answer. > > Our governance page here: > https://cwiki.apache.org/confluence/display/CASSANDRA/Cassandra+Project+Governance > > Felt kind of unclear on what committer votes were binding for or not, and > what pmc votes were binding for or not. I tweaked heading sizes, added > horizontal dividing lines, and then changed the following line: > Committer <https://projects.apache.org/committee.html?cassandra> votes are > considered “binding > <https://www.apache.org/foundation/voting.html#binding-votes>” > To: > Committer <https://projects.apache.org/committee.html?cassandra> votes are > considered “binding > <https://www.apache.org/foundation/voting.html#binding-votes>” on code > > Which leads to a question for everyone. Currently, on that wiki doc we have: > How we vote as a community: > > Which previously just had 'Committer votes are considered "binding"'. Then > proceeded to talk about commit votes, CEP votes, and release votes. Committer > votes are _not_ binding on releases as that's a legal thing with the ASF; > have to be a PMC member for that to be binding which was sort of alluded to > with the text: > Consensus: min 3 PMC +1, no PMC -1. These votes are neither Simple majority > nor Super majority. Vetos are to be accompanied with (technical) rationale. > > For code commits we've considered committers binding. > > For CEP's.... ? Commit-bit binding? Or pmc binding? I think it's the former > (committers) since that's the way the wiki reads, but I think we've been > acting as PMC only binding for CEP's? I think clarifying this explicitly > (i.e. in all 3 sections for code, CEP, and release, specify who's binding) > would be good. > > What do we think? Is there a consensus and I'm the only one that's confused? >
