> On Sat, Jul 17, 2010 at 5:32 PM, Richard L. Hamilton > <rlha...@smart.net> wrote: > > > Until Indiana was self-hosting, they built on SXCE. > > AFAIK, once Indiana was self-hosting, they built on > it. > > > > Some (no doubt not all) would like to stay in sync > with what they're > > building on, as much as possible. > > > > How well will a community distro be able to do > that? > > Either we heave-ho and try to chase IPS, wherever > that leads us, and > try to stay as much as possible in lock step with > where their > installation/BE/build architecture takes us; or we > realize that we > DON'T have to play the NIH game that started IPS and > we fill in the > gaps with some other enviroment/packaging. Chris B. > and the OSUnix > gang made a heck of a run getting things to build > with another system > and packaging. Is it critical that we try to create > an upgrade path > from 2009.11?
For a kewl desktop or appliance, probably not. For contributing, it would help to stay reasonably in sync. If IPS or something tied into it (the be* commands?) were involved, it might be necessary. A packaging system is a packaging system. IPS is nobody's favorite, but that's better than arguing the merits of rpm vs deb vs BSD ports vs ... > Look, there's source for several consolidations. > It'll build on at > east some version of a recent build for the moment -- > or at least > some version of it does. Nexenta used a completely > different > packaging system and they managed to track and > incorporate changes > until recently. I know you've been running down the > information gap > and reverse engineering things -- imagine if you > actually had help > with that aspect. > > There's one edge that we have that Oracle does not, > by the way. > Consider for a moment the /contrib repository. Ask > yourself if Oracle > has the resources to maintain a large port tree. Sun > had ramped up > resources and infrastructure for enabling the growth > of such a tree -- > including internally managing many ports in house -- > and Oracle seems > to have taken a 180 on this approach. That's, to me, > a significant > difference. The bottleneck seems to be approvers - people who can bless something to go from /pending to /contrib. Fixing that would go a long way toward increasing the package count (which is a pretty shallow metric, but it's one everyone seems to look for). -- This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org