Re: gconfd and multiple machines

[seth vidal <skvidal phy duke edu>]

> This should not be as hard as one might think, because a) dbus can do
> the networking and kerberos aspect for us b) the implementation of
> actually putting stuff on disk, etc. can be the same as for the
> per-user daemon c) with dbus the conversion to/from wire protocol is
> also the same for the per-user and the network-wide daemon.  I have a
> good idea from doing dbus how to write one daemon that can work for
> both systemwide and per-user.
> 
> Still, it's a fair bit of work.
> 

How much evolution has dbus seen? Before something like that could be
trusted a fairly good sized code audit would need to be done. 


-sv