Re: Getting ready to release 1.10.0

[Christopher <ct...@gmail.com>]

I reviewed the PR. Looks like it's headed in a good direction. I
wouldn't bother creating a 1.10 branch. We can do that after we tag a
release and do the normal branch cleanup.

On Mon, Dec 23, 2019 at 9:37 AM Mike Miller <mm...@apache.org> wrote:
>
> I am working on a 1.10 branch which is the 1.9 branch + JDK8 +
> updates/ignores to appease the modernizer.  @Christopher
> <ct...@gmail.com> I took your changes to PropertyType from your branch
> and they look pretty good.  I think we are able to update a lot in that
> class since it is fairly self contained.  A lot of the updates to
> configuration classes though, I will just annotate with @SuppressModernizer
> since we can't drop Mock in a 1.10 release.  I just started the work but
> was wondering how I should push the branch.  Do you want to review it on a
> PR against 1.9 that I can push in a 1.10 branch? Or want me to just create
> a 1.10 branch in the main repo whenever its ready and send out a link for
> folks to review?
>
> On Wed, Dec 11, 2019 at 1:34 PM Christopher <ct...@apache.org> wrote:
>
> > I think the main hold up is that I had committed to doing the JDK8
> > changes we discussed to get a 1.10, and I haven't had time to finish
> > following through (too focused on 2.1 development and other tasks). I
> > had thought there was interest in a 1.10, but for me, I'd be just as
> > happy aborting that plan and sticking with maintaining the 1.9 line.
> > Then, neither I nor anybody else has to bother doing any JDK8 changes
> > or dealing with the modernizer plugin. I'm reluctant to simply ignore
> > the plugin, since it does highlight quality issues that we don't want
> > to become worsened over time if we simply ignore it. I'd much prefer
> > to abort the 1.10 plan and stick to supporting 1.9 as the LTS release,
> > and doing a 1.9.4. I don't want to do both because I don't want to
> > support both.
> >
> > If we drop the 1.10 plans, I think we could do a 1.9.4 release pretty
> > quickly (I could prepare a test release candidate pretty much right
> > away to star the ball rolling, and an RC1 within a week). But, I don't
> > want to go against what the community agreed to, by vote, without
> > another vote (which I don't want to champion).
> >
> > On Wed, Dec 11, 2019 at 10:15 AM Adam Lerman <al...@gmail.com> wrote:
> > >
> > > I agree with Mike that one or the other should happen. If the hold up is
> > > the modernizer plugin, and we know those issues are addressed in 2.0+,
> > > maybe we consider ignoring those issues in the 1.10 line.
> > >
> > > On Wed, Dec 11, 2019 at 9:57 AM Mike Miller <mm...@apache.org> wrote:
> > >
> > > > If the demand is still there for a 1.10, I am willing to help out.
> > But in
> > > > the meantime, we could still release a 1.9.4, regardless of whether we
> > have
> > > > a 1.10 or not.  The last bug fix for 1.9 was in April and there have
> > been
> > > > some fixes and improvements that users are waiting on.
> > > >
> > > > On Tue, Nov 12, 2019 at 8:24 PM Michael Wall <mj...@gmail.com> wrote:
> > > >
> > > > > Thanks Christopher.
> > > > >
> > > > > On Tue, Nov 12, 2019 at 3:48 PM Christopher <ct...@apache.org>
> > wrote:
> > > > >
> > > > > > It was pointed out to me that some of the problems I had using the
> > > > > > modernizer-maven-plugin could be alleviated by adding granular
> > > > > > exceptions in the modernizer config. I'll see if I can make that
> > > > > > happen.
> > > > > >
> > > > > > On Fri, Nov 8, 2019 at 7:14 PM Christopher <ct...@apache.org>
> > > > wrote:
> > > > > > >
> > > > > > > So, one of the problems I've run into migrating the 1.x code to
> > JDK8
> > > > > > > is that we still have Mock in the API. That was removed in 2.x.
> > > > > > > However, in 1.9, MockConfiguration extends a non-public API,
> > > > > > > AccumuloConfiguration, which uses non-public Guava types for
> > > > > > > Predicates. Code-modernization/quality checks performed by the
> > > > > > > modernizer-maven-plugin catch the use of Guava's Predicate. I'm
> > still
> > > > > > > looking at this, to see if I can work around it without breaking
> > > > > > > anything, but it's a bit frustrating, especially since the right
> > fix
> > > > > > > (removal of Mock) was already done in 2.x.
> > > > > > >
> > > > > > > I'll revisit next week after a long weekend. In the meantime, if
> > > > > > > anybody is having second thoughts about a 1.10 release, or
> > opinions
> > > > > > > about what to do, feel free to express them here. One option is
> > to
> > > > > > > simply disable the modernizer-maven-plugin and ignore those
> > checks...
> > > > > > > but I don't really like the idea of disabling one of our tools
> > that
> > > > > > > does quality checks (even if these are very minor quality items).
> > > > > > >
> > > > > > > On Thu, Nov 7, 2019 at 4:20 PM Christopher <ct...@apache.org>
> > > > > wrote:
> > > > > > > >
> > > > > > > > As agreed in the recent [VOTE] thread, we will be releasing a
> > > > 1.10.0
> > > > > > > > that bumps the minimum runtime Java version to 8.
> > > > > > > >
> > > > > > > > I am beginning to work on getting the branch (currently still
> > named
> > > > > > > > 1.9) ready for release in accordance with this plan. As such, I
> > > > will
> > > > > > > > be preparing a pull request that bumps the Java version and
> > > > resolves
> > > > > > > > any errors generated by our plugins which do quality checks
> > > > > > > > (specifically modernizer, but also some compiler warnings).
> > > > > > > >
> > > > > > > > However, there are still a few other outstanding issues that
> > were
> > > > > > > > previously labeled for 1.9, which have not yet been resolved.
> > At
> > > > > least
> > > > > > > > one of these was labeled a "blocker". These are:
> > > > > > > >
> > > > > >
> > > > >
> > > >
> > https://github.com/apache/accumulo/issues?q=is%3Aopen+project%3Aapache%2Faccumulo%2F8
> > > > > > > >
> > > > > > > > If anybody is able to work on these, it will be very helpful in
> > > > > > > > getting the 1.10.0 ready for a release vote.
> > > > > > > >
> > > > > > > > Thanks,
> > > > > > > >
> > > > > > > > Christopher
> > > > > >
> > > > >
> > > >
> >