The way I interpret https://semver.org/, deprecated functionality should be
removed in a major release.

Users should be able to update to a new minor release without having to
worry that any of their code will break. It's certainly more convenient for
us to mark as deprecated and then remove whenever after the next major.
It's unrealistic though that users will have kept track of what was
deprecated before the previous major release came around and take that into
account every time they upgrade a minor. In essence we'd expect the user to
keep a list of what previously deprecated functionality they are still
using and treat every minor upgrade like a major until that list is empty.
That would be the only way for them to avoid these potential, very negative
surprises.

That said in this case it sounds like we already have broken large parts of
this API. Regardless of how we chose to interpret semver we should be able
to at least remove the parts of the API that's already broken.

On Tue, Apr 2, 2019 at 6:55 PM Jinmei Liao <jil...@pivotal.io> wrote:

> +1 since it's been deprecated pre geode 1.0, and I have no idea what its
> purpose.
>
> On Tue, Apr 2, 2019 at 5:20 PM Jacob Barrett <jbarr...@pivotal.io> wrote:
>
> >
> >
> > > On Apr 2, 2019, at 5:04 PM, Dan Smith <dsm...@pivotal.io> wrote:
> > >
> > > I think the best course of action is probably to remove it entirely.
> > > However, this does bring up a couple of questions:
> > >
> > > 1) Is it ok in general to remove deprecated API in a non X.0 release
> (eg
> > > geode 1.10 instead of geode 2.0)?
> >
> > I think a deprecated API is no longer API after a major version
> increment.
> > The deprecated symbols can be removed anytime after the major release
> with
> > merely a notice in the release notes. I think its unrealistic to expect
> all
> > deprecated symbols be removed for the x.0 release.
> >
> > > 2) How about in this case, when this feature has been deprecated for so
> > > long, and may or may not work?
> >
> > Given that this API was effectively deprecated in version 0 of Geode and
> > was never part of the release API in 1 it is fair game now. Time should
> not
> > be a factor, just major version number.
> >
> > -Jake
> >
> >
>
> --
> Cheers
>
> Jinmei
>

Reply via email to