Re: ideas on gconf-usage on multiple computers

[Ryan Lovett <ryan OCF Berkeley EDU>]

Argh, didn't want this to hit the list, but oh well... :)

On Wed, Apr 14, 2004 at 10:44:28AM +0200, Daniel Holbach wrote:
> I'm not quite sure, if I should use dbus, but maybe I'm mistaken about
> its use and features.

I had suggested dbus since, at first glance, you appear to want to write
a functional equivalent of gconf while the free desktop community is
potentially interested in gconf over dbus. (which should also be a
functional equivalent of gconf)

> Could you please briefly describe, how to use it in this environment:
> 
> Box A is the server running my service, awaiting clients (all kind of
> apps) on boxes B and C to update or tell them their preferences. The
> service tries to interpret the queries based on a ruleset provided by a
> config file and access XML-files, plain text stuff or databases to get
> or save data.

At present, to my understanding, this is exactly how gconf works in an
NFS environment. A user logs in to A where gconfd via ORBit writes out a
CORBA ior. When the user logs into both B and C, gconfd on B and C pick
up the ior from the user's home directory and send preference updates
back to A. This holds true if the gconf local lock mechanism is not set
and ORBit is configured to permit connections from other hosts.

gconf currently stores preferences in XML, though it supports multiple
backends.

> I'd love to use dbus, since I think it's a good thing in general, but I
> can't see how to do it at the moment.

This person looks like he's got a head start.

   http://people.imendio.com/micke/gconf-dbus/

Of course, if I'm misunderstanding your design, then all of this is
moot.

Ryan