On Thu, Oct 16, 2008 at 04:11:17PM +0100, Darren J Moffat wrote: > Thought that does bring up an interesting side issue about how the > nameservice versions of these will get fully populated if we don't > deliver the same content on SPARC and x86. Also shouldn't at least some > of that profile apply to SPARC too (not this case though).
One possible answer is that prof_attr/exec_attr/auth_attr(4) should mostly reside locally, with only overrides in the name service, "so who cares?". Another is that for Sun-delivered content we should never use i.rbac, nor any IPS action that might supplant i.rbac, that instead we should always integrate all prof_attr/exec_attr/auth_attr(4) entries into ON's master copy of those. Either way, not this case. Nico --
