accumulo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sean Busbey <bus...@cloudera.com>
Subject Re: compatibility check between 1.5.x and 1.6.0?
Date Wed, 23 Apr 2014 16:32:30 GMT
I suspect it was after the extended conversation about dumping mock in
favor of always using MAC.


On Wed, Apr 23, 2014 at 11:28 AM, Keith Turner <keith@deenlo.com> wrote:

> On Wed, Apr 23, 2014 at 12:03 PM, Sean Busbey <busbey@cloudera.com> wrote:
>
> > The default japi configurations were set to ignore the mock package.
> >
>
>
> Oh, thats not good.   I probably did that, but I do not remember why.
>
>
> >
> >
> > On Wed, Apr 23, 2014 at 10:58 AM, Keith Turner <keith@deenlo.com> wrote:
> >
> > > On Wed, Apr 23, 2014 at 10:47 AM, Sean Busbey <busbey@cloudera.com>
> > wrote:
> > >
> > > > Okay, I think all of these incompatibilities are things that should
> not
> > > > have been in the public API in the first place.
> > > >
> > > > * client.admin.SecurityOperationsImpl
> > > > * client.admin.TableOperationsImpl
> > > > * client.admin.InstanceOparationsImpl
> > > > * client.mock.MockShell
> > > > * client.mock.MockTabletLocator
> > > >
> > > > These changes are due to refactorings outside of the public API
> leaking
> > > > into classes within the client that handle implementation. For these
> > > > things, I think we should fix them to not be in the public API and
> just
> > > > include an apology in the release notes.
> > > >
> > >
> > >
> > > +1
> > >
> > > I remember seeing the .*Impl classes and ignoring them because they
> were
> > > Impl, but I shouldn't have because they are not in an impl package.   I
> > > don't remember seeing anything about mock, I must have glossed over
> that.
> > >
> > >
> > > >
> > > > Any objections before I make a ticket and put up a patch? Do we need
> a
> > > vote
> > > > about breaking the API, or is calling out that we're going to do that
> > in
> > > > the RC vote sufficient?
> > > >
> > > > (the other findings appear to be the japi compliance checker not
> > > > recognizing a method moving up a class hierarchy)
> > > >
> > > >
> > > >
> > > > On Wed, Apr 23, 2014 at 1:27 AM, Sean Busbey <busbey@cloudera.com>
> > > wrote:
> > > >
> > > > > Here's the current reports, built with japi-compliance-checker
> 1.3.6
> > > > >
> > > > > http://people.apache.org/~busbey/compat_reports/accumulo/
> > > > >
> > > > > executed with:
> > > > >
> > > > > for version in 1.5.0 1.5.1 1.5.2; do japi-compliance-checker
> > > > > -skip-deprecated -old japi-accumulo-${version}.xml -new
> > > > > japi-accumulo-1.6.xml -l accumulo; done
> > > > >
> > > > > with the config xmls like what's in the repo, but set to look in
my
> > > maven
> > > > > repo and to not skip mock.
> > > > >
> > > > > I'll triage the 1.5.0 changes tomorrow
> > > > >
> > > > >
> > > > >
> > > > > On Tue, Apr 22, 2014 at 9:04 PM, Eric Newton <
> eric.newton@gmail.com
> > > > >wrote:
> > > > >
> > > > >> I believe Keith did a mechanical/manual analysis and brought
back
> a
> > > > couple
> > > > >> of methods/classes.
> > > > >>
> > > > >> -Eric
> > > > >>
> > > > >>
> > > > >> On Tue, Apr 22, 2014 at 8:46 PM, Sean Busbey <busbey@cloudera.com
> >
> > > > wrote:
> > > > >>
> > > > >> > Has anyone done a compatibility check for the public API
between
> > > 1.5.x
> > > > >> and
> > > > >> > 1.6.0?
> > > > >> >
> > > > >> > While looking into ACCUMULO-2722 I noticed some changes
that
> might
> > > be
> > > > >> > problematic in client.mock and was wondering if anyone did
a
> > larger
> > > > >> sweep
> > > > >> > already.
> > > > >> >
> > > > >> > --
> > > > >> > Sean
> > > > >> >
> > > > >>
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Sean
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > Sean
> > > >
> > >
> >
> >
> >
> > --
> > Sean
> >
>



-- 
Sean

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message