[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [ecf-dev] ECF project and Tycho

----- Original Message -----
> From: "Markus Alexander Kuppe" <ecf-dev_eclipse.org@xxxxxxxxxxx>
> To: ecf-dev@xxxxxxxxxxx
> Sent: Wednesday, November 5, 2014 3:54:42 PM
> Subject: Re: [ecf-dev] ECF project and Tycho
> On 05.11.2014 14:27, Aleksandar Kurtakov wrote:
> > Well AFAIK, every committer to a project is supposed to have access
> > to LTS shared code repo for the project if the project is added to
> > the program. My personal take on LTS is that the program is still
> > taking shape but in order to show it's potential it needs some
> > consistency and providing the full stack in predictable way (the
> > reason for this mail). How successful it can or will be is an open
> > question. I still try to find my way in it and fill the gaps that
> > overlap with other things that we are involved in.
> OK, so once a project is in, its committers are enabled to fix bugs for
> LTS consumers?
> > - We look for collaboration. Second part is that as someone involved
> > in Eclipse platform - being able to provide additional(separate)
> > aggregator build that can build ecf together with everything else
> > verifying in one go that a change (e.g. migrating to newer apache
> > httpcomponents version) becomes a possibility if both parts use same
> > build system while in the current case it would be too much effort
> > which would alleviate the benefits. Last, but not least - the
> > conversion(if any?) would have to happen in the most beneficial way
> > for ecf project, so I appreciate that there is someone looking at
> > it.
> This is exactly opposite to ECF's current build which is motivated by
> OSGi's modularity concept. Why rebuild component/bundle A, if one of its
> dependencies (B) changes? If B indeed changed in incompatible ways,
> semantic versioning will tell. Otherwise, just re-aggregate and re-run
> (pre-built) tests. Even Tycho can consume pre-built bundles from a (p2)
> repository.

The case I'm looking for is why shouldn't I be able to provide a way to rebuild A and B in the same go to verify a change that has been done in C (which they both depend on) and how A and B interact with this change. Strictly devel environment.

> BTW in the broader OSGi community, Tycho sees less adoption compared to
> bnd. Thus, why replace Buckminster with Tycho, when ECF's consumers are
> primarily the OSGi community at large? Following the argument to ease
> contributions by adopting the most prominent build system, ECF would
> have to switch to bnd instead.

I can't not comment on that as I don't have the info needed about bnd. Is there such an agreement between ECF committers? Was it discussed somewhere (please point me to it if yes)? 
And the main question for me - is there interest in collaboration on Tycho based build system?

Alexander Kurtakov
Red Hat Eclipse team

> M.
> _______________________________________________
> ecf-dev mailing list
> ecf-dev@xxxxxxxxxxx
> To change your delivery options, retrieve your password, or unsubscribe from
> this list, visit
> https://dev.eclipse.org/mailman/listinfo/ecf-dev