Yes I meant the staging repo.

Take a look at the link I sent before about Apache spark release
announcement to get some tips how to prepare release for vote.

On Wednesday, December 11, 2013, Ankit Kumar wrote:

> Hi Henry,
>
> Can you explain what is the staging area? Is it the staging repo location i
> already shared in the vote thread earlier.
>
> I need to understand what extra steps i must perform before sending again
> the vote thread.
>
> Regards
> Ankit
> On Dec 11, 2013 3:15 PM, "Matt Franklin" <[email protected]> wrote:
>
> > Keep in mind the first Apache release is the hardest.   It gets a lot
> > easier from here.
> >
> >
> > On Tue, Dec 10, 2013 at 8:03 PM, Henry Saputra <[email protected]
> > >wrote:
> >
> > > Thanks guys, I guess we could continue with release VOTE using the
> > staging
> > > repo.
> > >
> > > I have set the svn pub sub for MetaModel to push our release artifacts
> > > later to the dist location.
> > >
> > > Ankit, as the RE for this release, can resend the [VOTE] thread with
> > > RC files in the staging area.
> > >
> > >
> > > - Henry
> > >
> > > On Sat, Dec 7, 2013 at 4:18 AM, Matt Franklin <
> [email protected]>
> > > wrote:
> > > > On Fri, Dec 6, 2013 at 3:18 PM, Kasper Sørensen <
> > > > [email protected]> wrote:
> > > >
> > > >> Hi all,
> > > >>
> > > >> Recent vote of the release didn't go too well because of variying
> > > opinions
> > > >> on where the release candidate artifacts should be provided. With
> this
> > > >> discussion I want to get an impression of what people need when they
> > > review
> > > >> a release, and what this means for the requirements and wishes of
> the
> > > >> actual release artifact location.
> > > >>
> > > >> It seems the norm in Apache projects is that artifacts are manually
> > > >> uploaded by the Release Engineer to somewhere in his personal
> > > >> people.apache.org web space. So some people prefer this approach.
> The
> > > >> upside here is that the release engineer can arrange the location so
> > > that
> > > >> it is as humanly consumable as possible. (Proponents of this
> approach
> > > may
> > > >> add more upsides that I may not know about.)
> > > >>
> > > >
> > > > Uploading to people.apache.org is actually an older practice that
> has
> > > been
> > > > deprecated. There is now a SVN Pub/Sub system with a staging and dist
> > > > section that we can ask INFRA to configure for us.
> > > >
> > > >
> > > >
> > > >>
> > > >> However, we're using Maven for the release packaging, signing and
> > > >> verification, and uploading to a Maven staging repository on
> > > >> repository.apache.org. This means that we can automize the upload
> > > >> completely by the existing Maven release process. The location that
> > the
> > > >> Release Engineer would publish would ultimately be a small Maven
> > > >> repository. This has the advantage that if you want to review a
> > release
> > > >> candidate from the perspective of an existing project that is
> > depending
> > > on
> > > >> MetaModel, you could simply consume the repository in your project
> and
> > > >> update your dependency version number. That way you could use the
> > > release
> > > >> candidate with very few changes and be able to build, test and run
> > other
> > > >> projects that depend on it. It also means that the location of the
> > > release
> > > >> candidate artifacts is not on people.apache.org, and that for
> > instance
> > > the
> > > >> source zip file will not be in the root of the upload location, but
> > in:
> > > >>
> > > >>
> > >
> >
> [root]/org/apache/metamodel/MetaModel/[version]/MetaModel-[version]-source-release.zip.
> > > >> That location is obviously less easily accessible if you just get
> the
> > > >> repository URL.
> > > >>
> > > >
> > > > We should deploy to repository.apache.org.  It ensures are artifacts
>

Reply via email to