Mailing-List: contact dev-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hbase.apache.org
MIME-Version: 1.0
In-Reply-To: 
 <CADcMMgFbH39WKYFkSieT7PK9CNaeDh3eV4PP8fFs+C2oN1zeLQ@mail.gmail.com>
References: 
 <CAGHyZ6+5aXUgdcPwirHivetHt_mvDbN8fzo-LAgzidZp8ddy9A@mail.gmail.com>
 <CANZa=Gt9H_0G7JawyEnwa_FNeTFw9Z-sqT9qfC36dyvQACf7jg@mail.gmail.com>
 <CADcMMgFbH39WKYFkSieT7PK9CNaeDh3eV4PP8fFs+C2oN1zeLQ@mail.gmail.com>
From: Misty Stanley-Jones <mstanleyjones@cloudera.com>
Date: Thu, 3 Dec 2015 15:21:35 +1000
Message-ID: 
 <CANqQv0uFOe+gpFMH38QRwF4RBORPo18tPPN=k5SEXG2w_cwE7Q@mail.gmail.com>
Subject: Re: [DISCUSS] Moving the HBase site to use a stand-alone git repo
To: "dev@hbase.apache.org" <dev@hbase.apache.org>
Content-Type: multipart/alternative; boundary=001a1148ef10af04290525f791f7

--001a1148ef10af04290525f791f7
Content-Type: text/plain; charset=UTF-8

The Jenkins job would checkout the main repo's master branch, run 'mvn
clean site site:stage' to generate the site, docs, and APIdocs. Then it
would checkout the asf-site branch of the hbase-site repo (or whetever it
is called) and commit the newly-generated target/stage/* to it and push.
Does that make sense?

On Thu, Dec 3, 2015 at 6:33 AM, Stack <stack@duboce.net> wrote:

> Good by me. Interested in the answers to Nicks questions too.
> St.Ack
>
> On Wed, Dec 2, 2015 at 11:07 AM, Nick Dimiduk <ndimiduk@apache.org> wrote:
>
> > +1 in theory. How will this work with integration of javadoc into the
> site?
> > How will RM's manage integrating site docs into their releases?
> >
> > On Wed, Dec 2, 2015 at 8:56 AM, Sean Busbey <busbey@cloudera.com> wrote:
> >
> > > Hi Folks!
> > >
> > > You may recall the occasional emails dev@ gets from a Jenkins job
> Misty
> > > set
> > > up to make updating the website easier for us. They're titled "HBase
> > > Generate Website" and they give a series of steps any committer can run
> > to
> > > push the changes live.
> > >
> > > Misty has been investigating automating this entirely[1], so that once
> > > updates land in the master source branch the website just updates. IMO,
> > > this would go a long way to improving how consistently updates make it
> to
> > > our primary public-facing presence.
> > >
> > > During our conversation with INFRA (on the jira[1] and in a
> infra@apache
> > > thread), the consensus seems to be that having an automated non-human
> > > process push to a repo that doesn't contain source that might lead to a
> > > release is acceptable. In contrast, such non-human pushing to our main
> > repo
> > > (even if just to the asf-site branch) is seen as higher risk that would
> > > require a policy decision.
> > >
> > > Is everyone (especially PMCs) fine with us moving our site to a
> different
> > > repository?
> > >
> > > Presumably something like hbase-site. The expectation is that in almost
> > all
> > > cases folks won't need to checkout or track this remote since the
> > automated
> > > job will be pushing rendered updates for us.
> > >
> > >
> > > [1]: https://issues.apache.org/jira/browse/INFRA-10722
> > >
> > > --
> > > Sean
> > >
> >
>

--001a1148ef10af04290525f791f7--