As Robbie stated, 5.16.0 should go out as is so option 1 is the best one to me. We need to cut off non blocking changes and get the release out as it has been a long time coming. As long as it is runtime compatible with JDK 8+ and runs fine on JDK 11 then that will work.
In 5.17.0 we can remove LevelDB and support JDK 11 builds. Hopefully we can get this release out faster as it would be nice to be able to finally use JDK 11 to build. We have the option in 5.17.0 to still target JDK 8 or we could just target 11+, but either way JDK 11 needs to work to build. On Wed, May 6, 2020 at 4:45 AM Robbie Gemmell <[email protected]> wrote: > As with the other times this was discussed, I think 1) makes most > sense still. If you think the bits are in a releaseable state now, > then it seems sensible to proceed with the existing stable base rather > than expanding and likely delaying the release yet again. Releasing > gets the finished changes people have been waiting for out, and you > can start fresh on big changes for the next version at the front of a > release cycle rather than lumping them in at the end. The next set of > changes can then arrive on their own however soon they are ready for > release, in a smaller targetted release. > > On Wed, 6 May 2020 at 06:31, Jean-Baptiste Onofre <[email protected]> wrote: > > > > Hi all, > > > > I would like to release ActiveMQ 5.15.13 and 5.16.0. > > > > They include fixes affecting several users (on http proxy, on xstream > CPU use, etc) and couple of improvements. > > > > About 5.16.0, we have two options: > > 1. I’m releasing as it is: it’s JDK 11 compliant at runtime, but the > build is not fully JDK 11. > > 2. I’m removing leveldb, scala, and update the branch to fully build > with JDK 11. I can postpone this for 5.17.x with JMS 2.x support (already > started). > > > > Thoughts ? > > > > Regards > > JB >
