Thanks for taking this on. Krisztián can confirm the details (or point you
to where this is documented), but based on past patch releases, I believe
you would make a `maint-4.0.x` branch off of the existing `release-4.0.0`
branch, cherry-pick the commits associated with the JIRAs tagged for 4.0.1
(I believe there are utility scripts to help with this), and run the
release script that bumps the versions.

Neal


On Fri, May 14, 2021 at 12:32 PM Jorge Cardoso Leitão <
[email protected]> wrote:

> Just to make sure: the goal is to cherry-pick all changes targeted for
> 4.0.1 into a branch and release from there? If that is the case, then I
> will create a branch and start cherry-picking the changes in order they
> were merged in master.
>
> I see 5 issues on the list still open. I subscribed to them and will be
> cherry-picking them as they get merged.
>
> On the Rust side; we can either bump 4.1.0 or cherry-pick for a 4.0.1. I
> suggest 4.0.1 to keep parity as we recently agreed, but let me know if
> others disagree.
>
> [1] https://github.com/apache/arrow-rs/pull/289
>
>
>
> On Thu, May 13, 2021 at 5:54 PM Neal Richardson <
> [email protected]>
> wrote:
>
> > Thanks, Jorge!
> >
> > If anyone else has bugfixes that they'd like included in a patch release,
> > please tag them with the 4.0.1 Fix Version. Perhaps we can do a roundup
> and
> > start a vote early next week?
> >
> > Neal
> >
> > On Thu, May 13, 2021 at 8:20 AM Wes McKinney <[email protected]>
> wrote:
> >
> > > Addressing these accumulated issues in a patch release sounds like a
> > > good idea to me.
> > >
> > > On Wed, May 12, 2021 at 6:18 PM Jorge Cardoso Leitão
> > > <[email protected]> wrote:
> > > >
> > > > I agree. Segfaults are not nice.
> > > >
> > > > I can take it. I would possibly need some guidance.
> > > >
> > > > Best,
> > > > Jorge
> > > >
> > > >
> > > >
> > > > On Thu, May 13, 2021 at 12:52 AM Neal Richardson <
> > > > [email protected]> wrote:
> > > >
> > > > > Hi,
> > > > > As discussed at the biweekly sync call, I wanted to gauge interest
> in
> > > doing
> > > > > a 4.0.1 patch release.
> > > > >
> > > > > There currently are 14 issues in JIRA tagged with 4.0.1 [1]. There
> > are
> > > 3
> > > > > segfaults, including one that a cuDF maintainer raised yesterday
> [2]
> > in
> > > > > requesting a patch release.
> > > > >
> > > > > I don't want to bias the discussion by giving my opinion (yet). I
> > will
> > > say
> > > > > that the question is whether someone (or multiple people) wants to
> > > step up
> > > > > and drive a release--if releases were costless, this would be much
> > > > > different. We did decide to allow for a simpler patch release
> process
> > > > > (source vote only, not on binary artifacts), so this could be a
> test
> > > for
> > > > > whether that does simplify matters and/or lets us better distribute
> > the
> > > > > work of producing binary artifacts.
> > > > >
> > > > > Any thoughts--especially from those who could/would be release
> > manager?
> > > > >
> > > > > Thanks,
> > > > > Neal
> > > > >
> > > > >
> > > > > [1]:
> > > > >
> > > > >
> > >
> >
> https://issues.apache.org/jira/issues/?jql=project%20%3D%20ARROW%20AND%20fixVersion%20%3D%204.0.1
> > > > > [2]:
> > > > >
> > > > >
> > >
> >
> https://lists.apache.org/thread.html/r9444ad73db11fe071a59ef5c084c6d1a02d26d55d23320ea8e0a67ae%40%3Cdev.arrow.apache.org%3E
> > > > >
> > >
> >
>

Reply via email to