Indeed -- the release would come with backward compatibility guarantee
within the given major version (modulo known exceptions: @Internal,
@Experimental, util package, etc.), all according to the standard semantic
versioning [1]. Also, see pending BEAM-1345 [2] that calls for audit and
sprinkling of annotations on parts that are not ready to be frozen. Anybody
should feel free to send PRs to keep things subject-to-change, if
appropriate.

Going forward, we'll have to add add automated testing for compatibility --
tracked by BEAM-1131 [3], and I believe Jason is on it.

[1] http://semver.org/
[2] https://issues.apache.org/jira/browse/BEAM-1345
[3] https://issues.apache.org/jira/browse/BEAM-1131

On Thu, May 4, 2017 at 12:53 PM, Thomas Weise <[email protected]>
wrote:

> Since the release would come with backward compatibility promise (?),
> creation of the release branch would also affect what happens on master
> from that point onwards. It would be good to provide some guidance on that?
>
> --
> sent from mobile
> On May 4, 2017 12:08 PM, "Davor Bonaci" <[email protected]> wrote:
>
> > We've been working on the first stable release for a while now -- we've
> > been tracking blocking issues [1], resolved many of them, organized a
> > hackathon [2], and improved testing coverage in various areas.
> >
> > I think we are nearly done -- there are a few more blocking issues left
> in
> > JIRA, and probably a few more that we'll discover as we go along. That
> > said, I think it is time to discuss the process for getting this release
> > finalized.
> >
> > I'd like to propose the following (tweaked) process for this special
> > release:
> >
> > * Create a release branch, and start building release candidates *now*
> > This would accelerate branch creation compared to the normal process, but
> > would separate the first stable release from other development on the
> > master branch. This yields to stability and avoids unnecessary churn.
> >
> > * Community-driven acceptance criteria
> > This would be an added step where everyone can recommend necessary
> > conditions for acceptance of the release candidate. Those conditions
> would
> > be *double-checked* manually, in addition to having all automated tests
> > pass. This could include important scenarios that we want to be really
> sure
> > about, e.g., WordCount example works on a given runner, TextIO can read
> > from HDFS, and we'd jointly validate those scenarios.
> >
> > * Vote
> > As usual.
> >
> > Time-wise, I think it would be really awesome to make this release
> coincide
> > with the ApacheCon conference that starts in about 2 weeks (on May
> 16th). I
> > think this would be a great goal -- and I'll do my part to make this
> > possible.
> >
> > Finally, we haven't formally closed on the version designation (see
> > separate thread).
> >
> > Thoughts? If there are no objections, I'll start this process soon, and
> we
> > can adjust as we go.
> >
> > Thanks!
> >
> > Davor
> >
> > [1] https://issues.apache.org/jira/issues/?jql=project%20%
> > 3D%20BEAM%20AND%20fixVersion%20%3D%20%22First%20stable%
> 20release%22%20AND%
> > 20resolution%20%3D%20Unresolved%20ORDER%20BY%20due%20ASC%2C%20priority%
> > 20DESC%2C%20created%20ASC
> > [2]
> > https://docs.google.com/document/d/1UKC2R_9FkSdMVTz2nt2sIW18KoLbIu6w0aj9
> > bwSSPiw/
> >
>

Reply via email to