KDEREVIEW: nowplaying dataengine and applet for plasma

[Ian Monroe]

On Jan 30, 2008 6:27 AM, Kevin Krammer <kevin.krammer at gmx.at> wrote:
> On Tuesday 29 January 2008, Ian Monroe wrote:
>
> > I find MPRIS to be a fine standard, it has the methods it needs to get
> > the job done. I'm hardly a DBus expert, where I've heard most of the
> > criticism coming from it seems. So for the folks criticizing it, just
> > join the MPRIS mailing list[*] and voice your concerns and suggestion.
> > Because in about 6 months to a year (with releases from Amarok,
> > VideoLAN and others), it is going to be the defacto standard.
>
> I tried that almost a year ago, doesn't seem to work.
>
> It's a bit hard to understand why there is so much resistance to appropriately
> group the methods into three interface. Anyone using a code generator like
> qdbusxml2cpp will need to do a lot of dummy method implementations when all
> objects have the same interface but only care about a few methods.

You just have to pass an extra option to either the cmake command or
to qdbusxml2cpp to change the classname. You can see the code in
kdemultimedia/dragonplayer/src/dbus, I didn't have to do anything too
strange (though I kind of messed up my naming scheme, that's my
fault).

Admittedly its a pain (especially since the cmake command doesn't seem
to work /quite/ right), but not something to go change an interface
over.

> It also messes up the discoverability. Any user using tools like qdbus will
> see all methods on each node and probably wonder why certain methods return a
> D-Bus error when being called.

I haven't noticed this problem. That said, I'll ping the list and ask
why we're doing it this way.

Ian