On Tue, Dec 23, 2008 at 10:02:16AM -0800, Stephen Hahn wrote: > * Nicolas Williams <[email protected]> [2008-12-23 08:02]: > > So... is it worthwhile? > > It will be, eventually--don't toss your prototype. The point of > providing the less blunt/more desired tools first is to force > migration of the packages we have as part of the distribution. The > larger set of packages that exist outside the distribution might very > well need a service like yours, but I'd prefer to extract known common > operations first, as we did with drivers and users.
If we have time to do manual migration of SVR4 package scripting from across the WOS, then indeed, my automatic migration prototytpe will be primarily useful only for third parties. My hope was to allow the manual migration to progress independently of any release deadlines should deadline pressures be too much. I definitely agree w.r.t. common operations. RBAC profiles, and registrations for most plugin frameworks should be implemented as native IPS actions. Many/most of the CAS in ONNV seem to exist for upgrade reasons only, not for RBAC profile editing, plugin framework registrations, etc. That means that if there's no Solaris 10/Nevada->OpenSolaris upgrade path [and no plan for such a feature] then most of these CAS can be thrown out altogether. OTOH, if anyone wants to build such an upgrade facility, then those CAS can live again through the SVR4-scripting->IPS prototype. I've not looked carefully at what all the post* scripts do, but a sample seems to indicate many are there just to deal with driver registration. Nico -- _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
