[Freeipa-devel] Ready to release?

[Nathaniel McCallum]

On Fri, 2014-07-04 at 17:34 +0200, Petr Spacek wrote:
> On 4.7.2014 17:20, Petr Viktorin wrote:
> > On 07/04/2014 04:57 PM, Martin Kosek wrote:
> >> Hello developers!
> >>
> >> I would like to thank everyone for the hard work during the last weeks,
> >> when finishing the FreeIPA 4.0 release, I saw many last stabilization
> >> fixes in DNS, OTP, ACIs, upgrade and Web UI areas.
> >>
> >> The last major work that is still not pushed is the CA management tool.
> >> Unfortunately, the final development and review did not go so well and
> >> we still do not have final patches.
> >>
> >> Given that this is a major piece of work (>50 patches) and given that
> >> the reviewer is on a leave, I am considering moving the feature to next
> >> release - FreeIPA 4.1 which is short and would still make it to Fedora
> >> 21. I would not like to stall 4.0 any more, I would like to offer all
> >> the work we have done for wider testing as soon as possible.
> >>
> >> Thoughts? Any other last patches you would like to add to add to 4.0 GA?
> >> There is still little time, but remember, git tag hammer is ready to fire.
> >
> > I'm testing these two:
> > - mbasti-0098-100, DNS tests
> > - mkosek-478, Spec bump
> >
> > Then there's pvoborni-695 with only a functional ACK
> >
> > The CA management is still up for discussion, but it's more and more likely
> 
> We have deferred 'full' DNSSEC support already so it is not unprecedented.
> 
> I have one proposal (valid only if we decide to defer CA management):
> 
> Maybe we could release 4.1 in few weeks when DNSSEC and CA management are done 
> & properly reviewed.
> 
> That would allow us to follow 'release early, release often' lore. In my 
> opinion, smaller release = safer.
> 
> Of course, it would mean shifting current version numbers from 4.x to 4.(x+1), 
> sorry Martin! :-)

I like this plan. 4.1 could cover just DNSSEC, CA management and 4.0
fixes. The original plan for 4.1 could become 4.2. I don't think the
change in version numbers would imply a change in overall timing.

I completely agree that smaller releases are better. :)

Nathaniel