Nobody's touched this for a week, but I think the conversation is important enough that it shouldn't die. I'll tackle these in version order.
1.3 - I agree with releasing a source only tag of 1.3.7 and calling it end-of-life. The changes are there, throwing them away seems silly (and rude?). That said, I'm not volunteering to be the one to put it together, so I feel a bit guilty suggesting that somebody else do it. 1.4 - There are a dozen bugs already marked as fixed. There are 5 more issues marked as fix for 1.4.4 that are yet unresolved. There are 15 issues marked as affecting 1.4.3 that are still open, some of which I would really like to see make their way into a 1.4.4. Some of those issues look fairly major and could use a few more eyes on them. Some time in June seems reasonable. 1.5 - So far these tickets look like mostly documentation fixes, but a few of them look pretty important. ACCUMULO-1449 and ACCUMULO-1456 were the first to catch my eye. That said, as 1.5 gets run on more systems I'm sure more bugs will crop up. Personally, I wouldn't plan on releasing anything until the community starts asking for it. 1.6 - Feature freeze in 4 months would be... September? That gives plenty of time to run test and integration and still get something out by the end of the year. I think this is the release that people were talking about rolling to Java 7 and Hadoop 2, both of which will require more than the usual amount of testing, in my opinion. Six months after 1.5 would be mid-November, which I'm not confident in, but it's a nice target to shoot for. The major worry would be that if the timeline slips past that, then there won't be a release until sometime in 2014, given the tendency for activity to slow down around the holidays. Thank you for starting the discussion, Christopher. I hope that other people chime in as well. Mike On Sat, May 25, 2013 at 1:56 PM, Christopher <[email protected]> wrote: > I did a quick check through the JIRA to see if there were any > outstanding issues on old versions. It seems we've done a pretty good > job of keeping things up-to-date. > > I especially wanted to see if there would be any reason to maintain > unreleased working branches for these older versions if we were to > switch to git. > > Here's the summary pages of the latest unreleased versions of each line: > > 1.3.7: > https://issues.apache.org/jira/browse/ACCUMULO/fixforversion/12321844 > 1.4.4: > https://issues.apache.org/jira/browse/ACCUMULO/fixforversion/12324151 > 1.5.1: > https://issues.apache.org/jira/browse/ACCUMULO/fixforversion/12324399 > 1.6.0: > https://issues.apache.org/jira/browse/ACCUMULO/fixforversion/12322468 > > I took the liberty of putting tentative release dates in JIRA for the > unreleased versions, except 1.3.7. > > I think we should shut down 1.3.7 as end-of-life, either by releasing > a quick source-only tag with the additional two fixes since 1.3.6 or > throw them away and leave 1.3.6 as the final 1.3 tag. Either way, I > think we should remove its working branch in svn, and drop continuous > development support for it. Notice, there are only two minor fixes in > 1.3.7 from 1.3.6. > > I think we should release 1.4.4 sometime within the next month, so we > can focus on supporting 1.5 as the main stable line, and new features > in 1.6. > > I anticipated a 1.5.1 release in about 2 months. > > I think we should anticipate a 1.6 release 6 months from the last > release, if not sooner. This would mean a feature freeze in about 4 > months. > > -- > Christopher L Tubbs II > http://gravatar.com/ctubbsii >
