Idea for 3.3 ( was: Moving pop, imap and smtp koslaves from kdebase to kdepim )

Bo Thorsen bo at
Thu Jan 8 10:30:10 GMT 2004

On Thursday 08 January 2004 10:56, Helio Chissini de Castro wrote:
> Hi
> Instead of every time decide if an io slave must changes place as we
> have a lot of others that can be moved ( moving the current mentioned
> to kde-pim fastly leads someone propose move fish:// ftp:// to
> kdenetwork and tar/zip slaves to kdeutils, makes sense !! ), why not
> create KDEIO module ?
> This will really make sense, since every io slaves could be moved to
> this module, and even the newcomers ioslaves from nonbeta could fit
> easily in the module, without bother to decide if will bloat kdebase,
> or belongs to another place.
> From my packager point of view, will be a lot easier to manage small
> packages, and from the point of release KDE admin, will be easier to
> just enable/disable the ioslave subdir on compilation.
> Now, from the point of view of kdepim developers, i think that will be
> a lot easier update de kdeio module, than launch the inbetween releases
> on kdepim module without do a huge conflicts in packagers database.
> And for a developer point of view, this is enable new developers adopt
> more and more ioslaves to their applications, since now they become
> visible and easy to manage and not force the developer to install
> another kde module just to use and test

Best reason I can think of: Without such a module, absolutely nothing will 
work in KDE. Hence, it's as important as kdelibs and might as well be in 
there. I hate the idea of "you need to install these 47 different things 
before you go here". kdelibs is fine as a requirement, but anything more 
is bad. (kdeartwork btw now has a compile time dependency on kdebase.)

It's completely wrong to talk about not bloating a package just because 
it's split in two packages - bloat is about unnecessary stuff, not about 
module size.

I originally had an idea about being able to run kdepim apps without 
kdebase, but too many things are needed on runtime for this to be able to 
happen. And moving these slaves this discussion was about would not fix 
that problem - it would only be about giving pim people more control over 
the slaves they are almost exclusive users of.

My bigger vision would be to kill kdelibs and instead make a kdecore that 
would hold everything needed to run a kde application - for example in 
gnome - and let kdebase be for setting up the desktop environment. This 
would make sense to me. But then people would probably bitch about the 
size of such a module (after which they would install both kdelibs and 
base, proving they're wrong).

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: signature
URL: <>

More information about the kde-core-devel mailing list