devicenotifier with qgw

Jacopo De Simoi wilderkde at
Sat Sep 19 15:36:32 CEST 2009

On Saturday 19 September 2009 14:50:50 Marco Martin wrote:
> On Saturday 19 September 2009, Jacopo De Simoi wrote:
> > Dear Plasma-devs
> >   Giulio and I are currently porting the device notifier to use qgw, the
> > delegate stuff was really too rigid to implement decent animations and a
> > few features (e.g. busy widget on top of the item being (un)mounted). Of
> > course we'd like to use the shiny new itemBackground; I'd like to ask you
> > guys what  do you think about the following modifications:
> >
> > 1) we need to activate the items on hover, but with the itemBackground
> > animation delay, hoverEvent is not good anymore to track that. I can see
> > two solutons: -  itemBackground should send some signal when its animation
> > finishes, something like targetReached(qgi *) where the pointer would be
> > null if the target was not an item. -  itemBackground should make publicly
> > available the animation time, so that we can animate accordingly
> > fadein/fadeout in each item.
> you mean open the items with a mouse over? like doing a click?

I mean show further information (free disk space) and the left action (mount/unmount/eject). 
Since now there is a visual delay in the selection, it doesn't look good to use the hoverevents anymore, and we need to sync with the animation of itemBackground

> >
> > 2) we need a way to destroy the hover when the mouse leaves all items; i
> > propose this solution: - implement a method ItemBackground::hideRequest
> > (qgi* item) which does the following: if (current target != item) return;
> > else trigger a singleshot which would in turn trigger a setTarget(0) unless
> > setTarget is called in the meantime.
> you can keep it and just hide it destroying and creating it every time is not 
> a big gain.
> also, if you just call hide now it will fade out with an animation so i don't 
> think new methods are needed

The only concern I have in doing this is the fact that the signals leave/enter are not guaranteed to be in the right order (or are they?)

> > What do you think about that?
> > Thanks
> >
> >   Jacopo
> >
> > _______________________________________________
> > Plasma-devel mailing list
> > Plasma-devel at
> >

More information about the Plasma-devel mailing list