[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: ktouchpadenabler moved to kdereview
From:       Michael Jansen <kde () michael-jansen ! biz>
Date:       2012-01-05 1:08:42
Message-ID: 1751347.1XTHNrJhhd () gambit
[Download RAW message or body]

On Thursday, January 05, 2012 12:17:33 AM Albert Astals Cid wrote:
> El Dimecres, 4 de gener de 2012, a les 23:40:26, David Faure va escriure:
> > On Wednesday 04 January 2012 18:51:44 Albert Astals Cid wrote:
> > > El Dimecres, 4 de gener de 2012, a les 01:53:13, Christoph Feck va
> 
> escriure:
> > > > On Wednesday 04 January 2012 00:28:11 Albert Astals Cid wrote:
> > > > > My little kded daemon that listens to XF86XK_TouchpadToggle and
> > > > > enables disables the touchpad accordingly has been moved to
> > > > > kdereview.
> > > > > 
> > > > > My plan is moving it to extragear, not really sure if -base or
> > > > > -utils.
> > > > > 
> > > > > The code doesn't have a kcm or any kind of configuration since
> > > > > it
> > > > > is designed to "just work".
> > > > > 
> > > > > I'd appreciate any review or suggestion over it.
> > > > 
> > > > I cannot test it because I have no touchpad, but if it is supposed
> > > > to
> > > > "just work" without any UI, I suggest to just add it to "khotkeys"
> > > > or
> > > > "kaccel" daemon (whichever of them is used for global shortcuts), so
> > > > that we do not filter global X11 keyboard events twice.
> > > 
> > > I don't really see any point in doing that, nothing can be shared
> > > between
> > > them and the existing ktouchpadenabler so instead of one simple codebase
> > > (166 lines with 20 of headers) you end up adding more complexity to
> > > existing programs (probably integrating the code in the existing
> > > programs
> > > would be more than 166 lines).
> > 
> > IMHO this isn't about the number of lines of code, but about the runtime
> > performance (how many process to wake up when pressing a key).
> 
> khotkeys is already a kded module, so there won't be no more processes
> waking up now than before by adding a new kded module.
> 
> > kglobalaccel seems quite suitable indeed, no?
> 
> It would, if Qt had a key for XF86XK_TouchpadToggle, as it doesn't i'd need
> to introduce a big "ignore all the workflow of kglobalaccel for this
> special key" since kglobalaccel only understands Qt keys (see
> KGlobalAccelImpl::grabKey).

And in the current design it is unsuitable. Because it listens for keys and 
forwards them via dbus. And forwarding it to itself via dbus is kinda ... 
deadlocking.

But i am about to fix that anyway to move the "Disable global shortcuts" 
shortcut from kwin to kglobalaccel. So i could help in porting this code too.

Do not consider that a 'It should be moved'. I know nothing about that stuff 
to chime in on that.

Mike

[prev in list] [next in list] [prev in thread] [next in thread]