Hi All.

`branch-1.9` is created to prepare Apache ORC 1.9.0 and
`main` branch is bumped to 2.0.0.

- Milestone for 1.9: https://github.com/apache/orc/milestone/10
- Milestone for 2.0: https://github.com/apache/orc/milestone/20

In general, Apache ORC releases have been aligned with
Apache Spark releases and maintenance cycles
with a longer support period.

- ORC 1.6.x <-> Spark 3.2.x (End-Of-Life)
- ORC 1.7.x <-> Spark 3.3.x
- ORC 1.8.x <-> Spark 3.4.x
- ORC 1.9.x <-> Spark 3.5.x

In addition to the planned on-going 1.7.x/1.8.x/1.9.x,
we can start Apache ORC 2.0 as a good opportunity
to reduce technical debt like removing Hadoop 2 and
supporting Java 21 LTS.

Of course, we may want to create ORC 1.10.x later from `branch-1.9`.
Apache ORC v2.x is a parallel stream that will not block anything of 1.x.
If needed, we will discuss it later after Apache ORC 1.9 release.

Lastly, for now, `main` branch keeps ORC File Format `v0.12`
for Apache ORC 2.0 because ORC File Format Versions
(v0.11, v0.12, UNSTABLE-PRE-2.0) are orthogonal from
Apache ORC library versions (1.6.x ~ 2.0.x).
Like Apache Parquet or Arrow, we may want to explicitly split repositories
at this time.

Please let us know your opinion about the new `main` branch version and
release schedule.
Also, feel free to create GitHub Issues for the milestone in order to do
planning together.

Thanks,
Dongjoon.

Reply via email to