Java bindings for dbus
hp at redhat.com
Mon Feb 14 18:09:14 PST 2005
On Mon, 2005-02-14 at 17:49 -0500, Jeff Morgan wrote:
> On Mon, 14 Feb 2005 14:12:46 -0500, Havoc Pennington <hp at redhat.com> wrote:
> > Longer term you could also think about doing a pure Java dbus client
> > implementation; I think this would only be a few thousand lines of Java
> > at most, but I could be wrong. Once we go 1.0 you could do this without
> > having to stay in sync with changes to the protocol.
> This is a very interesting idea. How stable would you say the protocol
> is at this time ?
"Almost stable" ;-)
It could still change, you would have to stay on top of it. But I don't
_expect_ any significant changes.
The only thing I'd add here: you'd be crazy to do a pure java
implementation without a pretty substantial test suite to verify its
correctness. The C unit tests already have code to generate a bunch of
messages to parse, so if you can chain that C code to your Java
implementation somehow you could get good testing without a lot of
If you can get said testing in the main "make check" I'll even fix up
the Java implementation whenever the tests break ;-)
More information about the dbus