KDE 3.2 and aRts

[George Staikos]

On Sunday 16 March 2003 08:52, Stefan Westerfeld wrote:
> I think its very convenient for media players and media applications to be
> able to depend on some basic functionality being always provided, at least
> in terms of APIs. This is why I am working hard on obsoleting aRts in the
> sense of having an own aRts process that always runs.

  Understandable.

> On the other hand, it should be fairly convenient for media players, web
> browsers (and so on), to be able to call some function in the code which
> plays a sample and/or sounds ; without checking at compile time whether the
> API is present or not. Thats why I'd like to keep an depedancy to the IDL
> compiler, for writing and implementing own components, and for binding them
> to programming languages.

  Yes that would be quite nice.  Conditional compiles are not fun.  I know 
this all too well.

> However, one could probably make having the implementation of aRts optional
> (whereas having the interface would still be mandatory), and applications
> can then check at runtime whether an artsd implementation exists, where
> they can always rely on that an aRts interface exists.
>
> Thus the lines
>
>   Arts::Synth_PLAY p;
>   if (p.isNull()) {
>     // handle missing component
>   }
>   else {
>     // works.
>   }
>
> could perform a run-time check. Still due to portability issues, I could
> imagine using glib would solve some issues in the aRts code, which are
> for instance in the select() code.

   I don't see how glib can do something there that cannot be done otherwise.

> Anyway, for KDE3.2 I think remaining backward compatible to KDE3.0 is a
> quite important goal, which I'd like to achieve.

   Of course, I think this is a requirement for KDE, not just a goal.

-- 

George Staikos