OT: klipper's misbeliefs [was: NEdit & KDE loss of selection problems]

[Joerg Fischer]

Scott Tringali wrote:

> That's to say, the dual (selection/clipboard) model works... when every 
> application is aware of the dual model.  But there are applications 
> which aren't, and the user might want to transfer text around between 
> them.  I think that's one of the problems klipper is trying to solve.

Well, unfortunately doing so they are creating problems, and out of
curiosity (and because KDE is quite popular) I'm trying to figure out
what happens.

The discovery that it is the disowning of the empty selection inside
nedit which causes klipper's interference is verified in
clipboardpoll.cpp, where they indeed check only for a change in
ownership of the selection, see

http://websvn.kde.org/trunk/KDE/kdebase/workspace/klipper/

The only relation to "prevent empty clipboard" I can see so far is in
klipper.cpp in line 856, but I don't understand why nedit loses the
selection thereafter :-(

Cheers,
Jörg