For the server, then it would seem reasonable to document the interface 
in terms of the SMF properties. Although this does make the the 
documentation Solaris-specific - I'm not very worried about that for the 
server though since our primary deployment platform for the server is 
Solaris.

For the client, I don't see what SMF service would be associated with 
the image itself, such that the cfg_cache is a cache of SMF properties. 
So I don't see how the name follows for an image (user or full).

Tom

Shawn Walker wrote:
> Tom Mueller (pkg-discuss) wrote:
>> One problem with making this a documented interface is the name of 
>> the file.  The file isn't a "cache" in that it isn't a quicker to 
>> access copy of data that is found elsewhere.  (Same with the 
>> client-side cfg_cache file, but there we have the pkg(1) command in 
>> front of it to do the editing.)
>
> Actually, it will be a cache though on OpenSolaris systems.  It is 
> planned that all of these values will be exposed via SMF and then 
> cached in cfg_cache; hence the name for both the client and the 
> server.  Look at docs/repository.txt in the gate for an explanation.
>

_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to