[petsc-dev] does next model mess up our histories

[Satish Balay]

On Thu, 2 Oct 2014, Karl Rupp wrote:

> On 10/02/2014 04:34 PM, Jed Brown wrote:
> > Satish Balay <balay at mcs.anl.gov> writes:
> > > I disagree on 'reset next' every few days as a fix to this problem. I
> > > suspect
> > > Jed will have a different objection for this reset..
> >
> > It makes it harder to follow changes on 'next' (including whether
> > certain features get fixed) and users can't simply "git pull".  I think
> > it's useful to have 'next' persist between releases (we could do more
> > frequent feature releases).
> 
> I second more frequent feature releases. Currently we release every ~12
> months. Releasing every 6 months is certainly within reach. It should also
> help us with finishing features up on a regular basis rather than having them
> in a half-done state for too long (of course I'm also referring to myself
> here).

As long as we don't wait for a feature - we should be able to do
frequent timely releases [master is always supporsed to be in a
releaseable state :)]

One issue is: we currently support only 1 previous release at a
time. For faster releases - would this cause grief? [wrt users
requiring frequent upgrades, and us requiring to support more
releases]

Satish