Thanks, that's useful advice. This is part of the effort to integrate that huge list of FOSS into Open Solaris managed by Kelly Nishimura (See [1] & [2]). RPE was given a section of the list, and Venky volunteered (along with other RPE engineers) to perform the integration. I don't think he has any particular affiliation with the product or community, and there is no "project team" as such. So I don't know if anyone ever intended (or has the time) to work with the community on such things as IPv6 support.
Lukas Rovensky is the RPE OSS manager, maybe he can add something? Lukas, are there are plans/resources to improve these FOSS products beyond integration? If not, we may be faced with a familiar but simple FOSS choice - integrate it as is (because we have no resources to improve it), or don't integrate it. [1] http://infoshare.sfbay/twiki/bin/view/Main/OpenSolarisCabinet [2] http://infoshare.sfbay/twiki/pub/Main/OpenSolarisCabinet/OS.1120.V60.ods James Carlson wrote: > [Removed unusual cc-list.] > > James Gates writes: > >>I'm not too sure what we can discuss here. As the community haven't >>really warmed to IPv6, I would have thought that lack of IPv6 support is >>the norm, not the exception. What would be the outcome of an ARC >>discussion - Would you insist that the community implement IPv6 support? >>Would integration be stalled until IPv6 support is enabled? > > > Regardless of what "the community" may think of IPv6 (and perhaps what > I think of it as well), it's actually required in some markets, > including with US government customers, so integrating components that > don't support it poses potentially serious problems with at least > Sun's marketing and sales. > > For that reason, the Solaris PAC had a "big rule" requiring IPv6 > conformance for all Solaris components that include IP-related > networking. The relevance of that (and perhaps all "big rules") is > somewhat unclear to me at this point, but the only 'reasonable' > exception was if the protocol being implemented was itself not defined > to work with IPv6. (E.g., ARP doesn't work with IPv6, so an > ARP-related project wouldn't be forced into inventing something that > doesn't exist.) > > Yes, I'm well aware that there are some groups and products that never > did IPv6 and possibly never will. Driving compliance across something > as big as Sun is not easy. > > I would at least offer _strong_ ARC advice that the project team > consider implementing that support and donating the results to the > community -- unless the application is just plain horrible, this > shouldn't really be hard to do -- and suggest that they get in touch > with Victor Nelson to find out more about the issues and risks that > may be caused by spotty IPv6 support. > > (Including, in some cases, the risk of losing our IPv6 Logo. Likely > not a risk with this one project, but certainly a cumulative risk over > time if we don't try to guard the commons.) > -- Jim Gates Sun Microsystems Nashua, USA http://sun.com/postgresql
