This issue, from my perspective, somewhat speaks to the "maturity model"
surrounding an open source community.

If NiFi is to be considered a "stable and mature" product, such that any
minor point release is considered supported and usable for production
purposes, then a 0.2.1 release should most definitely be created with the
hotfix applied.

Jumping from 0.2.0 (with known show-stopping issues) to 0.3.0 doesn't yet
speak to a mature product development cycle.  That's OK, so long as
everyone acknowledges that NiFi is currently an alpha or beta quality
product.

But if NiFi at 0.2.0 is supposed to be "mature enough for generalized use,"
issuing a 0.2.1 hotfix release would definitely be the right thing to do
and would accrue more OSS karma.

+1 to Mike Moser's recommendation for a 0.2.1 release, iff the community is
ready to acknowledge and support a product that is more than perpetual beta
quality.






On Thu, Jul 23, 2015 at 1:12 PM, Joe Skora <jsk...@gmail.com> wrote:

> @Joe Witt
>
> I think you hit the nail on the head with "Perhaps we should *consider *it
> ...".  What makes sense in each circumstance will depend on the severity of
> the problem and the difficulty to fix it, with other factors including team
> considerations figuring in as well.  The ecosystem needs an active user
> community and an active project team, so there must be a balance, but all
> things being equal I give a slight preference to user considerations.
>
> On Thu, Jul 23, 2015 at 11:45 AM, Joe Witt <joe.w...@gmail.com> wrote:
>
> > ~Mike Moser “I'm just asking for very public documentation (perhaps on
> > the Downloads page itself) that 0.2.0-incubating has a known issue
> > that may cause people to decide to not use it”
> >
> >
> > Understood and agree.  I’ll update the downloads page to reflect that.
> >
> >
> > ~Joe Skora “I like the idea of a 0.2.1 "hot fix” ...it keeps the
> > changes in that release focused on the fix minimizing the testing
> > demands and allowing for quicker deployment.”
> >
> >
> > Understood.  My thinking on 021 vs 030 was based more on the effort of
> > the community to put out a release but your and Mike’s perspective
> > sounds like it is more from the ‘user’.  We have to take a balanced
> > view but it should always err on the side of the user experience.  In
> > short, point taken.
> >
> >
> > Perhaps we should consider it that if we generate a ticket against a
> > release and call that ticket a blocker then we should consider a
> > hotfix.  https://issues.apache.org/jira/browse/NIFI-777
> >
> >
> > I’ll initiate an 021 release package and send it out for a vote
> > [caveat being my current Internet situation is quite sketchy].
> >
> >
> > Thanks
> >
> > Joe
> >
> > On Thu, Jul 23, 2015 at 7:51 AM, Joe Skora <jsk...@gmail.com> wrote:
> > > +1
> > >
> > > In general, I like the idea of a 0.2.1 "hot fix" for two reasons.
> First,
> > > it keeps the changes in that release focused on the fix minimizing the
> > > testing demands and allowing for quicker deployment.  Second, it
> prevents
> > > the need to rush the next release just to expedite delivery of a fix.
> > > There are scenarios where a "hot fix" may not be a good idea, but in
> the
> > > case of a show stopping issue it minimizes the time that the project
> can
> > be
> > > considered to in a broken state.
> > >
> > > My 2 cents, YMMV.
> > >
> > > Joe
> > >
> > > On Thu, Jul 23, 2015 at 9:46 AM, Michael Moser <moser...@gmail.com>
> > wrote:
> > >
> > >> My intent wasn't to debate the Semantic Versioning 2.0.0 scheme that
> > Apache
> > >> NiFi has adopted.  My question was more on the lines of what to
> release,
> > >> and I was suggesting a release for BOTH a 0.2.1 and 0.3.0.  I
> understand
> > >> this community has discussed the desired Git workflow in the past.  I
> > can't
> > >> remember the decision, but if I'm asking for a "hot fix" release from
> > the
> > >> Git flow perspective and the community decided to not follow that
> path,
> > >> then I accept that.
> > >>
> > >> I guess I'm just asking for very public documentation (perhaps on the
> > >> Downloads page itself) that 0.2.0-incubating has a known issue that
> may
> > >> cause people to decide to not use it.  It's certainly causing our
> > community
> > >> to fast track the 0.3.0 release.
> > >>
> > >> Thanks,
> > >> -- Mike
> > >>
> > >>
> > >> On Wed, Jul 22, 2015 at 2:39 PM, Joe Witt <joe.w...@gmail.com> wrote:
> > >>
> > >> > Mike
> > >> >
> > >> > I believe that assigns too much meaning to the versions and may not
> > >> reflect
> > >> > the versioning model we have.  We have a wiki page that outlines the
> > >> > version scheme.  The idea here is that the develop branch is
> > essentially
> > >> > always ready to be released.  We simply set the version of the
> develop
> > >> > branch to be based on what the tickets dictate as per our version
> > scheme.
> > >> >
> > >> > I do agree for major releases like 1.0.0 we have a need to discuss
> as
> > a
> > >> > community how long to support pre 1.0 as well as 1.x.  This is based
> > on
> > >> our
> > >> > published versioning model.
> > >> >
> > >> > Does this sound reasonable to you?
> > >> >
> > >> > Thanks
> > >> > Joe
> > >> > On Jul 22, 2015 10:55 AM, "Michael Moser" <moser...@gmail.com>
> wrote:
> > >> >
> > >> > > All,
> > >> > >
> > >> > > As a suggestion for the future, at least after NiFi reaches 1.0, I
> > >> think
> > >> > a
> > >> > > bug fix release such as a 0.2.1 would definitely happen in a
> > scenario
> > >> > like
> > >> > > this.  Even after 0.3.0 is available for download, if you don't
> > >> recommend
> > >> > > that people use 0.2.0 due to major bugs, then you should replace
> it
> > >> with
> > >> > > 0.2.1 on the Downloads page.
> > >> > >
> > >> > > -- Mike
> > >> > >
> > >> > >
> > >> > > On Tue, Jul 21, 2015 at 5:31 PM, Aldrin Piri <
> aldrinp...@gmail.com>
> > >> > wrote:
> > >> > >
> > >> > > > All,
> > >> > > >
> > >> > > > If there are no objections, I am volunteering to take release
> > manager
> > >> > > > responsibilities for the 0.3.0 release.
> > >> > > >
> > >> > > > This process will include releases for three components:
> > >> > > > * nifi-parent, version 1.0.1
> > >> > > > * nifi-nar-maven-plugin, version 1.0.2
> > >> > > > * nifi, version 0.3.0
> > >> > > >
> > >> > > > A ticket [1] has been created to track the release process.
> > >> > > >
> > >> > > > Thanks!
> > >> > > >
> > >> > > > [1] https://issues.apache.org/jira/browse/NIFI-778
> > >> > > >
> > >> > > >
> > >> > > >
> > >> > > > On Mon, Jul 20, 2015 at 7:37 AM, Mark Payne <
> marka...@hotmail.com
> > >
> > >> > > wrote:
> > >> > > >
> > >> > > > > Ryan,
> > >> > > > >
> > >> > > > > Yup, I'll make sure that one is in there.
> > >> > > > >
> > >> > > > > Does anyone else have any ticket that they want to ensure gets
> > out
> > >> > > today?
> > >> > > > >
> > >> > > > > Thanks
> > >> > > > > -Mark
> > >> > > > >
> > >> > > > > ----------------------------------------
> > >> > > > > > Date: Mon, 20 Jul 2015 10:33:14 -0400
> > >> > > > > > Subject: Re: nifi-0.3.0 release
> > >> > > > > > From: rhendrickson.w...@gmail.com
> > >> > > > > > To: dev@nifi.apache.org
> > >> > > > > >
> > >> > > > > > Will the Amazon AWS processors fix make this release?
> > >> > > > > >
> > >> > > > > > On Mon, Jul 20, 2015 at 10:29 AM, Mark Payne <
> > >> marka...@hotmail.com
> > >> > >
> > >> > > > > wrote:
> > >> > > > > >
> > >> > > > > >> Joe,
> > >> > > > > >>
> > >> > > > > >> Not a problem. I updated JIRA to show the tickets that will
> > go
> > >> > into
> > >> > > > > 0.3.0.
> > >> > > > > >>
> > >> > > > > >> Thanks
> > >> > > > > >> -Mark
> > >> > > > > >>
> > >> > > > > >> ----------------------------------------
> > >> > > > > >>> Date: Mon, 20 Jul 2015 10:10:20 -0400
> > >> > > > > >>> Subject: Re: nifi-0.3.0 release
> > >> > > > > >>> From: joe.w...@gmail.com
> > >> > > > > >>> To: dev@nifi.apache.org
> > >> > > > > >>>
> > >> > > > > >>> Sounds good. Would like to see the items that end up on
> the
> > >> > release
> > >> > > > > list
> > >> > > > > >>> in jira.
> > >> > > > > >>>
> > >> > > > > >>> Also we need to put together good release notes this time.
> > >> > > > > >>>
> > >> > > > > >>> Thanks
> > >> > > > > >>> Joe
> > >> > > > > >>> On Jul 20, 2015 10:02 AM, "Mark Payne" <
> > marka...@hotmail.com>
> > >> > > wrote:
> > >> > > > > >>>
> > >> > > > > >>>> Hello all,
> > >> > > > > >>>>
> > >> > > > > >>>> We have found an issue that is present in nifi 0.2.0 that
> > >> > results
> > >> > > in
> > >> > > > > new
> > >> > > > > >>>> accounts not automatically being requested as they should
> > >> when a
> > >> > > new
> > >> > > > > >> user
> > >> > > > > >>>> attempts to access an instance of NiFi.
> > >> > > > > >>>>
> > >> > > > > >>>> This is not a security issue but could cause some
> headaches
> > >> for
> > >> > > > > >> operators
> > >> > > > > >>>> who are running secure instances of NiFi, especially new
> > >> > > instances.
> > >> > > > > >>>>
> > >> > > > > >>>> Unless there are any objections, I would like to go ahead
> > and
> > >> > > start
> > >> > > > > >>>> putting together an 0.3.0 release that has this issue
> > >> addressed,
> > >> > > the
> > >> > > > > new
> > >> > > > > >>>> Flume Processors, and any other tickets that are complete
> > and
> > >> > > ready
> > >> > > > to
> > >> > > > > >> go.
> > >> > > > > >>>> All others will be moved to version 0.4.0.
> > >> > > > > >>>>
> > >> > > > > >>>> In the absence of any objections, we will probably have a
> > >> > package
> > >> > > > > built
> > >> > > > > >>>> and ready to be voted on today.
> > >> > > > > >>>>
> > >> > > > > >>>> Thank you
> > >> > > > > >>>> -Mark
> > >> > > > > >>>>
> > >> > > > > >>>>
> > >> > > > > >>
> > >> > > > > >>
> > >> > > > >
> > >> > > > >
> > >> > > >
> > >> > >
> > >> >
> > >>
> >
>

Reply via email to