A draft for a standard for desktop configuration

[Jamie McCracken]

Avery Pennarun wrote:

> The trick is just to design the protocol such that it's okay to have both
> getint() and getstring() work on the same (numeric) value.  You don't want
> the protocol *enforcing* data types, just *allowing* them as an
> optimization.

The Dbus variant can hold any type so essentially we dont have to worry 
about the type at all. I dont think Philip was trying to include extra 
type info in the protocol but merely expressing what that variant might 
contain. Its probably best to remove that from the spec to avoid confusion.

btw the variant is self describing (it has a type signature) so you 
would query that signature to work out the type

-- 
Mr Jamie McCracken
http://www.advogato.org/person/jamiemcc/