Joe Clarke <jcla...@cisco.com> wrote: > On 11/8/18 16:46, Martin Bjorklund wrote: > > Hi, > > > > After the session today, it seems to me that one fundamental > > requirement (or non-requirement) is missing. How much branching does > > the solution have to support? The current solution (6020/7950) and (if > > I understood Rob Shakir correctly) also openconfig have linear > > versioning *per module*. If we can get agreement on this requirement, > > I think it will be easier to find a solution. > > We discussed this in the DT. We didn't want to discuss levels of > branching in the requirements as that pointed to certain solutions.
All requirements point to certain solutions, in a sense. But if you say that branching is not a requirement on a solution, I'm happy with that. > The > point of requirement 1.4 was to say that the DT felt previous versions > of modules needed to support fixes without bringing in elements from head. I think this means that you require branching. But is this still the point of the "new 1.4" that was mentioned in the session? However, as Rob Shakir mentioned in the session, there are other ways to do fixes / enhacements than branching a single module. You can add new functionality with augment and make changes with deviations. /martin > We discussed branching to satisfy this requirement as well as protecting > new features with if-feature and using deviations. > > Joe > _______________________________________________ netmod mailing list netmod@ietf.org https://www.ietf.org/mailman/listinfo/netmod
