Common VFS: GKIO experiment

[Michael Pyne]

On Tuesday 07 December 2004 08:00 pm, Rahul Sundaram wrote:
> Hi
>
> > If you want to help me with developing GKIO, please
> > let me know (My time
> > to work on this is limited).
>
> gkio is a polarising name. choose a better one. we
> dont want to unnecessarily alienate other DE and WM
> developers

He asks for technical help and you tell him to change the name?

Anyways, to the original poster: You could perhaps start off by making a 
KIOSlave that interfaces with gnome-vfs (I'm not sure how the different main 
loops will work though).  But please don't assume that this will be a simple 
replacement once you're done.

KIO (the framework, not the KIOSlaves) is an essential part of KDE, and many 
KDE applications, both those included with KDE, and those that people write 
independently.  Even if you were to somehow get the gnome-vfs kioslave to 
work, it's not like you could then just plop it in and replace KIO.  If it 
were to happen for KDE 3 it would have to have the *exact same binary 
interface*, which if you were to try to do this would mean you've 
reimplemented KIO, poorly.

If you were to try this for KDE 4 (where binary compatibility can be broken), 
I still wouldn't support it at that time for technical reasons unless the 
changes that KDE 3 code would have to make to support it were slight, as 
we've already got a difficult enough porting job ahead of us as it is.

Please remember that there are KDE apps that rely on KIOSlaves for things that 
may not have exact equivalents in gnome-vfs.  KMail is the most notable 
example, and there are probably others.

I'm not trying to discourage you, but what you're asking about isn't at all 
like many of the other successful fd.o integration projects that KDE has 
implemented, and there are other much more achievable integration projects 
that could be pulled off now, like updating the KDE gstreamer bindings, or 
even submitting proposals for shared bookmark settings or the like.

Regards,
 - Michael Pyne