Norm Jacobs wrote:
> Brian Utterback wrote:
>>
>>
>> Norm Jacobs wrote:
>>
>>> orphaning that may occur. There are several components in SFW that
>>> already deliver root packages with RBAC entries.
>>
>> But only one (SUNWmkcdr) delivers nothing else. What do you see as
>> the problem here?
>>
> In this particular case, I doubt that you are going to cause any real
> harm if the SFW and ON bits are ever out of sync, but by splitting the
> delivery, you are all but guaranteeing that it will happen at some
> time. The RBAC bits for this in ON are almost certainly going to be
> forgotten. If they are in SFW in packaging related to the component,
> they more likely to be attended to. Aside from that, by packaging
> them in a root package in SFW, you can add/remove the RBAC data with
> the functionality. If you simply put it in ON and leave it up to
> SUNWcsr to install, you leave potentially orphaned entries in the
> RBAC data that may or may not become a problem down the road.
> Finally, it seems to me that if/when IPS provides a better answer for
> delivering file fragments, we will want those fragments in with the
> functionality that they are supporting. Just my $0.02
>
> -Norm
>
>
I think I have a very mild preference for the separate root package, for
the reasons listed above. It does seem to me that Darren's concerns
here about a mostly trivial root package are mostly aesthetic (an extra
root package).
But, I also believe that either way is OK, and that the argument is not
architectural in nature.
We do need IPS to address this somehow, going forward -- but that's not
a problem for this case.
-- Garrett
