roadmap

[Matthew Johnson]

On Mon Aug 11 13:48, Thiago Macieira wrote:
> Matthew Johnson wrote:
> >On Sat Aug 09 20:48, smog zer wrote:
> >> >> Feature negotiation has to be end-to-end, over the bus daemon.
> >> >> That's a huge problem because you would have to negotiate with all
> >> >> your peers.
> >> >
> >> > I hadn't even thought of that... does seem like a mess.
> >> >
> >> > There's a simple workaround in the case of DBUS_TYPE_FLOAT - either
> >> > send it as a double or send it as a byte array. If you're doing 3D
> >> > vertices over dbus that probably is not outrageously more low-level
> >> > than your app already is.
> >> >
> >> > Havoc
> >>
> >> Why care with negotiations, just add it now while the "iron is still
> >> hot" and dbus does not have that many wire protocol implementations,
> >> and then notify the current wire protocol library maintainers.
> >
> >It's interesting to note that of the three current implementations
> >(mine: java, alp's: .net and the reference one) libdbus is the only one
> >which does _not_ support float...
> 
> Have you implemented code to detect whether you can send floats to the 
> other side?
> 
> If you haven't, I'd argue it's an incomplete solution.

No, you have to turn them on or off (currently an environment variable).
I don't know what alp does.

Some sort of feature negotiation is something we should think about. I
can't imagine we will never want to add features so it will be
necessary at some point.

Matt

-- 
www.matthew.ath.cx
D-Bus Java
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.freedesktop.org/archives/dbus/attachments/20080811/d7f0b5f4/attachment.pgp