Thanks Tim - that makes sense, agreed on not carrying a dependency-named
branch. I've reworked things to the standard version-keyed release-line
model (same shape as Apache Spark's conventions):

- Renamed main-hudi-0x -> branch-0.4, so it's a proper 0.4 release line
rather than a "hudi" branch. GitHub auto-retargeted the open PRs.

- branch-0.4 is the line we'll cut 0.4.0 from and cherry-pick patch fixes
onto (0.4.1, 0.4.2, ...). It stays on Hudi 0.x on purpose, which keeps
supporting folks who haven't moved to Hudi 1.x.

- Repointed the two open 0.4.0 PRs to branch-0.4: the utilities-jar fix
(#841) and xtable-spark-runtime for Hudi 0.x (#843), and updated the Maven
CI / License Check triggers to run on branch-0.4.

Plan from here: land #841 and #843 on branch-0.4, get CI green, then cut
rc1 from branch-0.4.

Thanks,
Vinish

On Sat, Jul 11, 2026 08:14 AM, Tim Brown <[email protected]> wrote:

> I think that makes sense. I don't want us to make a hudi branch though. We
> should have a 0.4.X branch and then cherry pick any fixes to that release
> line as patch releases to continue to support folks that have not upgraded.
>
> -Tim
>
> On Thu, Jul 9, 2026 at 2:31 PM Vinish Reddy Pannala <
> [email protected]> wrote:
>
> > Hi all,
> >
> > Following up on the last update - a scope adjustment for 0.4.0 so we can
> > get the release out the door.
> >
> > We're going to decouple Hudi table format v9 support and the cross-format
> > indexing work from 0.4.0 and target them for 0.5.0. Both are actively in
> > progress (v9 for the Hudi target in #834, indexing in the separate
> > [DISCUSS] thread), but they need more time and I'd rather not keep
> holding
> > rc1 on them. Landing them together in 0.5.0 also lets us do indexing
> > properly, since v9 is what unblocks Hudi's metadata indexes for tables
> > written through XTable.
> >
> > To proceed with 0.4.0, the one additional item we'll try to land is
> > xtable-spark-runtime (targeting Hudi 0.x) - the new, thin, drop-in way to
> > run XTable inside an existing Spark job (see the separate "[DISCUSS]
> > Introduce xtable-spark-runtime" thread and #836). It's a much better
> > on-ramp than the standalone utilities jar.
> >
> > Even without v9, 0.4.0 is a strong release:
> > - Parquet source support (#768)
> > - Paimon support
> > - Delta Kernel conversion target (#801)
> > - A new way to run XTable via xtable-spark-runtime
> >
> > Plan: once xtable-spark-runtime is in and CI is green, I'll cut rc1 and
> > start the vote thread. v9 + indexing will headline 0.5.0.
> >
> > Please shout if you'd like anything else in 0.4.0, or have concerns about
> > deferring v9.
> >
> > Thanks,
> > Vinish
> >
> > On Tue, Jun 30, 2026 11:02 AM, Vinish Reddy Pannala <
> > [email protected]> wrote:
> >
> > > Hi all,
> > >
> > > A quick status update on 0.4.0:
> > >
> > > The Hudi 1.x upgrade (#772) is now merged, which was one of the key
> items
> > > we were holding rc1 on. The main remaining piece is Hudi table format
> v9
> > > support for the Hudi target, which is actively in progress. Beyond the
> > > upgrade itself, v9 unblocks Hudi's metadata index capabilities (e.g.
> > > indexes) for tables written from other formats through XTable, so it
> sets
> > > us up well for richer cross-format indexing in future releases.
> > >
> > > Once v9 lands I'll do a final pass over the remaining 0.4.0 PRs and
> > follow
> > > up on cutting rc1.
> > >
> > > Thanks,
> > > Vinish
> > >
> > > On Thu, Jun 25, 2026 03:36 PM, Vinish Reddy Pannala <
> > > [email protected]> wrote:
> > >
> > >> Thanks Tim for the review. I have addressed all the comments, please
> > >> review again when you get a chance.
> > >>
> > >> On Sat, Jun 6, 2026 at 1:45 PM Tim Brown <[email protected]>
> > wrote:
> > >>
> > >>> Thanks Vinish! I've added some blockers to #772, mainly around
> testing
> > >>> the
> > >>> completion time ordering.
> > >>>
> > >>> -Tim
> > >>>
> > >>> On Fri, Jun 5, 2026 at 5:27 PM Vinish Reddy Pannala <
> > >>> [email protected]> wrote:
> > >>>
> > >>> > Hi all,
> > >>> >
> > >>> > Status update on the PRs targeted for 0.4.0:
> > >>> >
> > >>> > - Delta Kernel conversion target (#801): merged.
> > >>> > https://github.com/apache/incubator-xtable/pull/801
> > >>> >
> > >>> > - Parquet incremental sync (#768): merged.
> > >>> > https://github.com/apache/incubator-xtable/pull/768
> > >>> >
> > >>> > - Upgrade Apache Hudi to 1.2.0 (#772): in progress. We are working
> > >>> through
> > >>> > the remaining review comments and a CI/test failure on the branch.
> > >>> > https://github.com/apache/incubator-xtable/pull/772
> > >>> >
> > >>> > Once #772 is green and merged, I'll cut rc1 and start the release
> > vote
> > >>> > thread. Reviews on #772 are welcome in the meantime.
> > >>> >
> > >>> > Thanks,
> > >>> > Vinish
> > >>> >
> > >>> > On Mon, Jun 01, 2026 02:16 PM, Vinish Reddy Pannala <
> > >>> > [email protected]> wrote:
> > >>> >
> > >>> > > Hi all,
> > >>> > >
> > >>> > > Hope everyone is doing well.
> > >>> > >
> > >>> > > We are planning to cut rc1 for the next Apache XTable
> (incubating)
> > >>> > release
> > >>> > > shortly. Before we do, we'd like to get the following pull
> requests
> > >>> in,
> > >>> > as
> > >>> > > they are the last ones targeted for this release:
> > >>> > >
> > >>> > > - Delta Kernel conversion target (#801):
> > >>> > > https://github.com/apache/incubator-xtable/pull/801
> > >>> > >
> > >>> > > - Upgrade Apache Hudi to 1.2.0 (#772):
> > >>> > > https://github.com/apache/incubator-xtable/pull/772
> > >>> > >
> > >>> > > - Parquet incremental sync (#768):
> > >>> > > https://github.com/apache/incubator-xtable/pull/768
> > >>> > >
> > >>> > > If you have some time, please take a look and share any review
> > >>> feedback
> > >>> > > over the next few days. Once these are merged and CI is green,
> > we'll
> > >>> go
> > >>> > > ahead and cut rc1 and start the release vote thread.
> > >>> > >
> > >>> > > Please reply on this thread if there is anything else you'd like
> to
> > >>> see
> > >>> > > included in this release.
> > >>> > >
> > >>> > > Thanks,
> > >>> > > Vinish
> > >>> > >
> > >>> >
> > >>>
> > >>
> >
>

Reply via email to