felix-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stuart McCulloch <mccu...@gmail.com>
Subject Re: [VOTE] Versioning after a failed release
Date Mon, 07 Feb 2011 19:56:32 GMT
On 7 February 2011 19:50, Felix Meschberger <fmeschbe@adobe.com> wrote:

> Hi,
>
> Just to shed some light from other projects:
>
> * In Jackrabbit a tag was created in anticipation of an
>   imminent release 2.2.3. A bug was discovered. The tag was removed
>   and the next release will be 2.2.4 (there will not be a 2.2.3).
>
> * In Tomcat they voted on a release and found issues (quality
>   wise) so decided to declare 7.0.7 broken. The version number was
>   not reused and the release currently under vote has the next
>   version 7.0.8.
>

which would actually be allowed under option A because it says "can reuse",
not "must" (unlike option B)

personally I think it's a bad idea to try to create a single rule to cover
all possibilities - just leave it to the discretion of the release manager


> Regards
> Felix
>
>
> Am Montag, den 07.02.2011, 10:56 +0000 schrieb Jeremy Hughes:
> > A (non-binding).
> >
> > FWIW: Stuart brought me off the fence. If the first 1.0.x release is
> > 1.0.5 then under semantic versioning guidelines this would mean 5
> > releases of the bundle at previous micro numbers have already passed
> > which might cause users some confusion, or cause them to ask for an
> > explanation. This confusion persists for ever more. The number of
> > people seeing this admittedly lesser kind of confusion will likely be
> > far greater than the people who use the RC artifacts from a staging
> > repo, who should know better.
> >
> > On 5 February 2011 12:50, Stuart McCulloch <mcculls@gmail.com> wrote:
> > > On 5 February 2011 12:21, Guillaume Nodet <gnodet@gmail.com> wrote:
> > >
> > >> As long has the release has not been approved, the tag does not match
> > >> an official release, so it can be freely deleted.
> > >>
> > >
> > > yep, that's what I meant - another point to consider is users might see
> > > 1.0.5 and think it's stable (as it's not 1.0.0) whereas in fact there
> could
> > > have been 5 staged versions just to sort out license / dependency
> issues and
> > > no actual code changes
> > >
> > > Once the release is voted, I think everyone agree the tag becomes
> immutable.
> > >>
> > >> FWIW, Git is much better as a tag really correspond to a moment in the
> > >> history, not a branch (which actually makes more sense if you think
> > >> about it).
> > >
> > >
> > > agreed, git is better in this regard - but it can be hard to understand
> at
> > > times :)
> > >
> > >
> > >> On Sat, Feb 5, 2011 at 11:04, Felix Meschberger <fmeschbe@adobe.com>
> > >> wrote:
> > >> > Hi,
> > >> >
> > >> > Am Samstag, den 05.02.2011, 09:52 +0000 schrieb Sahoo:
> > >> >> On Friday 04 February 2011 04:48 PM, Stuart McCulloch wrote:
> > >> >> > it is easy to retag releases in svn
> > >> >> >
> > >> >> >
> > >> >> What exactly do you mean by "retag releases in svn?" Rename an
> existing
> > >> >> tag or using the same tag name to tag a different snapshot of
the
> source
> > >> >> code base? Neither should be done in my IMHO.
> > >> >
> > >> > Agreed, both is far too easy ...
> > >> >
> > >> > Regards
> > >> > Felix
> > >> >
> > >> >
> > >>
> > >>
> > >>
> > >> --
> > >> Cheers,
> > >> Guillaume Nodet
> > >> ------------------------
> > >> Blog: http://gnodet.blogspot.com/
> > >> ------------------------
> > >> Open Source SOA
> > >> http://fusesource.com
> > >>
> > >
> > >
> > >
> > > --
> > > Cheers, Stuart
> > >
>

-- 
Cheers, Stuart

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