I think I'm for scenario 2, but I would want to at least one other 1.3.x
release if for nothing else to fix
the ODE dependencies. In my opinion 1.3.3 depends is not build on stable
release versions, and
if we're leaping to the 2.0 trunk it might take a few months to
stabilize it. So I think we'll need to support 1.3.x
for prob at least another 6 months for bug fixes?
>Rafal and I think that we don't really have the resources to maintain
two development branches and we should drop one
The Riftsaw team is ready to start pushing code/fixes back upstream..
--Kurt
Tammo van Lessen wrote:
Hi ODE devs,
on the IRC channel, we had in interesting discussion yesterday about the
current branch situation and I (have been :)) volunteered to raise this
issue here. Although /trunk has been communicated to be the successor of
the 1.x branch, development still goes on on 1.x so that we basically
have two development branches currently. This leads to the predicament
that each fix has to be created, reviewed and committed twice and makes
it difficult to track whether both branches are sync (i.e. all bug fixes
have been ported to the respective other branch. Rafal and I think that
we don't really have the resources to maintain two development branches
and we should drop one. The important question actually is, which one.
Scenario I: Discontinue and drop trunk and continue development 1.x
Scenario II: Discontinue 1.x and move on with trunk
I personally would really regret choosing Scenario I, since trunk has
couple of great new feature such as the new IL, OModel versioning,
extension activities, process contexts, etc., and that with good cause.
Not to mention the hard work that has been put into that branch. So it's
rather not an option.
Scenario II has IMHO less of such negative impact. We will still need to
maintain the 1.x branch for bug fixes of running 1.x instances, so we're
not dropping something, however, all new features should go into
trunk. In addition, it's rather impossible to stabilize and improve
trunk with respect to performance etc. if it is not released.
My suggestion would be to go for II, prepare a 1.3.4 and 2.0-beta2
release, declare 1.x to be a dead-end and move on with development on trunk.
What do you think?
Cheers,
Tammo
