On Sat, Jan 4, 2014 at 3:14 AM, Benedikt Ritter <brit...@apache.org> wrote:
> 2014/1/4 Henri Yandell <flame...@gmail.com> > > > On Fri, Jan 3, 2014 at 2:35 AM, Benedikt Ritter <brit...@apache.org> > > wrote: > > > > > Hello Hen > > > > > > > > > 2014/1/3 Henri Yandell <flame...@gmail.com> > > > > > > > > > > > I'm +1 for calendar defined minor releases btw. I think we should be > > > > releasing monthly, if anyone has the energy to deal with the RMing :) > > > > > > > > > > RMing is not such a big deal. I mean it's tedious, but it's okay, if > you > > > only have to to it once in a while. The problem is, that I don't see > that > > > much development activity. But let's see how the next weeks turn out. > > > > > > > From my point of view, I tend to freeze up on activity when an RM is in > the > > works. > > > What could be done is selecting issues we want to have in 3.3 and mark them > as such in Jira. > I want anything whose patch gets reviewed :) > > > > Sure we could develop in parallel etc, but it's mostly about the > > issue tracker and not developing lots of new features, so it's rare that > > something has a long active life on a separate branch. The last time I > > remember such was when I rewrote the text.translate stuff. > > > > Agreed, branching is a pain in svn but this my work out better after we > switch to git. > > Could be, though bear in mind I like merging in SVN so I'm saying the above from the project management side of lots of branches and doing clean up rather than the technicals of svn vs git. > > > > > So I'm holding off on doing too much while you're in RM mode. I think > we'd > > benefit from a scheduled release cycle if you're up to handling it. An > > intent that we release at the end of every 1, 2, 3, something months. If > > we're releasing often, we shouldn't get sidetracked on the packaging as > > we'll know we can fix it soon and there will have been less time for any > > oddities to show up. > > > > This sounds like you're not in the mood to RM yourself :-) I think I can > manage to roll out a RC in the schedule you proposed. But it would be even > better if we could rotate. > Really not in the mood to RM :) My personal situation is such that I don't have large pieces of time to commit, instead it's 30min each night, and my experience in trying to RM has been that that isn't enough. Ideally it should be zero effort, it should be a case of: Every commit generates release artifacts needed for a vote. Automated alarm sends "Should next Lang be released?" to commons-dev. People focus on things being good for a week or so, then start to mark +1. Release done. Perhaps something to focus on at the next ApacheCon if there was interest. Hen
