XMMS Second Generation - Comments.

[Havoc Pennington]

On Tue, Aug 26, 2003 at 08:04:15PM +0200, Tobias Rundström wrote:
> On tis, 2003-08-26 at 19:53, Havoc Pennington wrote:
> > FWIW, tutorial docs will hopefully be forthcoming for the GLib and Qt
> > wrappers when those are written. What you're using now is essentially
> > low-level backend so it only has the function reference stuff and no
> > overview.
> 
> Is this meant to be interpreted lowlevel API that we use today will not
> be stable and that we SHOULD use the glib wrapper later?

The lowlevel API and ABI will be frozen with 1.0, for now it certainly 
isn't stable (I just spent a few days collecting feedback from KDE
hackers that will result in some large changes for example).
See the README and your compile line containing
-DDBUS_API_SUBJECT_TO_CHANGE for the details ;-)
I also warned you guys about this on IRC a couple months ago.

The point about the lowlevel API is that it doesn't contain any
"convenience functionality" and isn't intended to just work; 
it has to be manually hooked up to a main loop, threads
implementation, your object instances, etc. - because it
doesn't contain those things, it's meant to work with GLib, Qt,
Python, or whatever, that do contain them.

> How is QT support progressing?

Nobody has been working on it so far, but there have been many
discussions about it here at Nove Hrady. So this should evolve over
time. Based on comments here most likely it won't be fully locked down
until KDE 4, if KDE chooses to adopt D-BUS in some way.

My presentation from here has a slide with what "adopt" could mean. 
KDE developers still have very diverse views on when/what should be
done with dbus, and in large part I think it depends on what code is
written and how well that code works.

I just put up the Nove Hrady presentations at:
 http://ometer.com/novehrady_freedesktop_pres
 http://ometer.com/novehrady_dbus_pres

The dbus one has at least one mistake Martin pointed out about using
SASL for encryption; I did use it for that (by having a
"COOKIES_WITH_ENCRYPTION" mechanism) but this is kind of bogus.

Havoc