ignite-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sergi Vladykin <sergi.vlady...@gmail.com>
Subject Re: [VOTE] Apache Ignite 1.0.0 RC3 release
Date Tue, 17 Mar 2015 15:49:10 GMT
+1 for RC3

Sergi

2015-03-16 0:19 GMT+03:00 Konstantin Boudnik <cos@apache.org>:

> On Sun, Mar 15, 2015 at 09:46AM, Branko ─îibej wrote:
> > On 15.03.2015 09:28, Konstantin Boudnik wrote:
> > > On Sun, Mar 15, 2015 at 07:55AM, Branko ─îibej wrote:
> > >> On 15.03.2015 07:15, Konstantin Boudnik wrote:
> > >>> +1
> > >>>
> > >>> I still recommend you not to include RC# into top-level directory
> name when
> > >>> you create the zip file. RC is a logical state of the archive - not
> its
> > >>> physical property. Also, earlier I commented on how the checksums
> will be
> > >>> screwed if you try to re-pack the archive.
> > >> Actually, if I see "Ignite ver. 1.0.0-rc3" printed on my screen when I
> > >> run a program, then I do expect the -rc3 tag to be part of the source
> > >> directory name. It is valid and reasonable practice to make a public
> > >> release candidate. We do that at least once for every 1.x.0 release at
> > >> Subversion:
> > >>
> > >>
> http://subversion.apache.org/docs/community-guide/releasing.html#release-numbering
> > >>
> > >> The process we use there is:
> > >>
> > >>   * Release 1.x.0-rc1 as a formal ASF release
> > >>   * Wait for a "soak period" (typically 4 weeks)
> > >>       o during this time, users can take the candidate for a spin and
> > >>         report any problems
> > >>       o we fix the problems, and if any of them are serious enough
> > >>         (e.g., requiring a public API change), we roll -rc2 etc. and
> > >>         restart the soak period
> > >>
> http://subversion.apache.org/docs/community-guide/svn-soak-management.png
> > >>   * Once we've decided that the release candidate is stable, we
> create a
> > >>     new release based on the same tag, with only one change: the -rc*
> > >>     tag is dropped from all version numbers (that's basically a
> one-line
> > >>     change in svn_version.h).
> > >>
> > >> We've found that this works extremely well to make .0 releases more
> stable.
> > > I actually do agree that the described procedure is a great way to
> > > stabilize a release. However, i was under the impression that the goal
> of
> > > this exercise is to produce 1.0 release that'd be ready for IPMC vote
> > > from release expectation stand-point, not technical merits of the
> project.
> > > If i am mistaken on that - then i want to withdraw my comment.
> >
> > I'll quote Dmitriy (with my emphasis):
> >
> > On 13.03.2015 08:25, Dmitriy Setrakyan wrote:
> > > I am resubmitting 1.0 RC3 for a vote having addressed comments from
> Cos and
> > > Brane. /_*Note that this is a real RC, and not a 1.0 yet.*_/ We would
> like to
> > > spend another week testing it before announcing a final 1.0 release.
> Any
> > > feedback from the community about the RC, in the mean time, will be
> much
> > > appreciated.
>
> You right - sorry guys: I guess my comment was off-based. I think I need to
> learn to read all over again ;(
>
> Thanks for the clarification, Brane!
>   Cos
>
>

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