Having the 25th as the branch creation date sounds good to me. Regarding HIVE-29038 — if there's a quick fix available, I’m fine proceeding with it. Once the branch is created, we can wait for the DN regressions to be addressed there. In the meantime, any other fixes that land can also be cherry-picked to the branch as needed.
Just to clarify: my suggestion is simply to go ahead with branch creation and start the release process in parallel. We can continue discussing with folks whether specific tickets should be part of this release or deferred to the next — the intent is not to override anyone’s opinion. We should only push the tag once everyone is aligned. –Ayush On Mon, 23 Jun 2025 at 13:26, Stamatis Zampetakis <zabe...@gmail.com> wrote: > > It's great to see that we are so close in cutting the next Hive release! > > My take as usual is to not hold off the release unless there are > serious issues, regressions, and legally blocking stuff. Anything that > falls into the category nice to have can go into the next release > assuming that it will follow-up in the next few months. > > Regarding the versioning scheme, I feel that the current content > deserves a major version bump (instead of 4.1.0 we should probably > consider 5.0.0). There have been many changes (~500 commits) since the > 4.0.1 and among them we definitely have breaking changes, behavior > changes, new features that are gonna be noticeable by end-users. The > JDK17 upgrade and the drop of support for 8 is one example, the > upgrade of Calcite from 1.25.0 to 1.33.0 highlighting affecting the > query optimizer is another one, and I am sure we can find many more. > > Best, > Stamatis > > On Mon, Jun 23, 2025 at 8:38 AM Denys Kuzmenko <dkuzme...@apache.org> wrote: > > > > RestCatalog is definitely not a blocker, however, would be good to reach an > > agreement on HIVE-29038 before releasing the new configs. > > For now, we can move forward with Iceberg-specific configs and handle > > unification in a follow-up release. > > WDYT? > > > > HIVE-28744 <https://issues.apache.org/jira/browse/HIVE-28744>: Fix the row > > count estimation for INT and BIGINT types > > is not a regression, but should improve row count estimates and performance > > > > Only HIVE-29030 and HIVE-29034 are regression-related items
