On Sun, Jun 5, 2016 at 9:25 AM, Dan Kenigsberg <danken(a)redhat.com> wrote:
On Sat, Jun 04, 2016 at 04:17:32PM +0300, Nir Soffer wrote:
> On Wed, Jun 1, 2016 at 4:21 PM, Nir Soffer <nsoffer(a)redhat.com> wrote:
> > On Wed, Jun 1, 2016 at 4:19 PM, Martin Sivak <msivak(a)redhat.com> wrote:
> >>> 1. master
> >>>
> >>> vdsm-4.19.0-201606011345.gitxxxyyy
> >>
> >> Ack and +1 to the idea, but I have one small comment. Isn't it usual
> >> in Fedora (for example) to use the following?
> >>
> >> vdsm-4.19.0-0.201606011345.gitxxxyyy
> >>
> >> Please note the zero in the release part (-0.something). The stable is
> >> then released as vdsm-4.19.0-1 keeping the version intact.
> >
> > Thanks for correcting me Martin, I omitted the release number mistake.
> >
> >>
> >> Martin
> >>
> >> On Wed, Jun 1, 2016 at 12:51 PM, Nir Soffer <nsoffer(a)redhat.com>
wrote:
> >>> Hi all,
> >>>
> >>> We are going to branch 4.0 today, and it is a good time to update our
> >>> versioning scheme.
> >>>
> >>> I suggest to use the standard ovirt versioning, use by most projects:
> >>>
> >>> 1. master
> >>>
> >>> vdsm-4.19.0-201606011345.gitxxxyyy
> >>>
> >>> 2. 4.0
> >>>
> >>> vdsm-4.18.1
> >>>
> >>> The important invariant is that any build from master is considered
newer
> >>> compare with the stable builds, since master always contain all stable
> >>> code, and new code.
> >>>
> >>> Second invariant, the most recent build from master is always newer
compared
> >>> with any other master build - the timestamp enforces this.
> >>>
> >>> Thoughts?
>
> Dan?
Yes, it's a good idea. I'd appreciate if it is implemented in such a way
that there is no need for an explicit commit to introduce a new version.
Currently it's done by a mere `git tag`.
But that's not a hard requirement. Feel free to have a "bump version"
commit like most other projects out there.
I think we can keep the current way we set the version, but replace
the serial number with a timestamp.
Nir