Andrea Pescetti wrote: > Kay Schenk wrote: >> would suggest given this >> clarification on what a "binding" vote means, those cast by PMC members, >> we need to definitely allow for a longer timeframe for release voting >> -- at >> least a week for a final RC candidate I would think. > > Release voting is not the moment for PMC members to learn how to build. > I would not encourage arbitrary extensions of the standard 72 hours for > voting (I would still be open to extend the period if someone requests > an extension during the vote and has a valid reason for it of course). > > Between the moment a Release Candidate is tagged and the moment the vote > finishes we already have a week or so, since building the release and > uploading it takes days. So people who want to vote based on the sources > can and should start building when we tag a release, and they will have > more time this way. > > So a process that could work is: we tag sources (say, RC3 sources), we > inform the dev list and we make the source package available. People can > start building straight away. When binary builds are ready and uploaded > (which takes days), we start the vote on that same source package plus > the corresponding binary builds. > > Regards, > Andrea. Andrea;
By my reading of the policy that would still mean that no binding vote could be cast by any pmc member that had not built the from the source. Whether we as a project like it or not or think it is right as far as the Apache Software Foundation is concerned the Release is the Source package and the Source package alone. The Policy document very clearly sets that fourth. I see two possible ways to deal with that. The first would be to apply to the Foundation for an exception that would allow the scenario you laid out in your earlier e-mail to allow binding votes to be cast specifying what they are voting for. The other would be for a separate but parallel vote on the Binary files. Regards Keith
signature.asc
Description: OpenPGP digital signature
