+1 to the second option of releasing 0.21.0 asap and postpone 0.22.0.
If we do end up finding bugs in 0.21.0, they can be fixed into 0.22.0
instead of creating a smaller release via 0.21.1.


On Wed, Mar 31, 2021 at 12:43 PM Jihoon Son <jihoon...@apache.org> wrote:

> Hi all,
>
> The 0.21.0 release has been blocked by several issues for a long time
> including 2 security vulnerabilities (CVE-2021-25646 and
> CVE-2021-26919 which are addressed in 0.20.1 and 0.20.2,
> respectively). Since we created the release branch for 0.21.0 on
> January 11th, it is almost the time to prepare the next release (the
> next branch cut is scheduled in April). If we stick to the schedule
> and prepare another release right after 0.21.0, it will not be great
> for our users because they won't have enough time to test out the
> 0.21.0 release. I think we have 2 options to avoid this issue:
>
> - Consolidating the 0.21.0 and 0.22.0 releases. We can create one
> release that will contain all commits that are originally scheduled to
> be released separately via 0.21.0 and 0.22.0.
> - Finalizing the 0.21.0 release quickly and postponing the 0.22.0
> release e.g., a month.
>
> To me, the second option seems more reasonable because of the below
> reasons.
>
> - The 0.21.0 release is almost ready. We can finish it quickly without
> much further preparation.
> - There are users waiting for the features introduced in 0.21.0.
> Consolidating 2 releases will make them wait longer.
> - There could be unknown bugs in the release branch. Consolidating
> releases will make testing harder since it will increase the scope of
> testing.
> - Consolidating releases will remove one option for users what release
> they can use. This can be problematic especially when some bugs are
> found in the new release.
>
> I'd like to hear what other people think.
>
> Thanks,
> Jihoon
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@druid.apache.org
> For additional commands, e-mail: dev-h...@druid.apache.org
>
>

Reply via email to