I agree with Ufuk. I'm uncertain if its a good idea to delay the 0.8.0 for these "last minute" streaming features. I have the impression that the current master is really stable and very well tested. If we now bring in some prototypes and last minute features, we probably end up discussing a 0.8.1 bugfix release in two weeks. I would prefer a 0.8.0 release asap because it has already been way to long with the issues unfixed in 0.7.0. Also, if the graduation is going to happen soon, we'll be unable to release for two weeks or so because INFRA has to transfer us to new namespace (mailing lists, website domain, release spaces, ... )
As a general lesson from this discussion, we still have to improve our communication to get a better mutual understanding of what we are working on. I'll create a wiki page today that contains a roadmap of the features we are working on. On Mon, Dec 8, 2014 at 1:28 PM, Ufuk Celebi <u...@apache.org> wrote: > On Mon, Dec 8, 2014 at 11:05 AM, Márton Balassi <balassi.mar...@gmail.com> > wrote: > > > Ok guys, then I also agree to skip 0.7.1 and go straight for 0.8.0. As > for > > the streaming side we would like to finish a couple of features (lambda 8 > > support, type handling rework, filesystems i/o support, at least once > fault > > tolerance prototype). I'm confident that we can get most of the things > that > > we really want in there done by the end of this week. > > > > > This sounds like a lot. I can understand that you want as much in a release > as possible though. ;) Would it make sense to reduce this list to the most > important changes and keep the rest for 0.8.1? >