Comment on global shortcuts security

Piotr Rak piotr.rak at
Tue Sep 25 09:28:20 PDT 2012

2012/9/25 Timothée Ravier <siosm99 at>:
> Le 25/09/2012 01:53, Bill Spitzak a écrit :
>> Keystrokes should be sent to the application first. Only if the
>> application refuses them should they be considered global shortcuts.
> According to me, the main goal is to _never_ have global shortcuts.
> That's why each applications should register the semantic keys they want
> to listen to.
> This is a distribution/desktop environment default configuration
> problem. I don't think there is a simple solution for this.
> The semantic approach means that only a special set of key combos will
> be available for applications to register.
> Those key combos will most likely correspond to non-printable characters
> as users would be annoyed by the double effect of one key.
> It could even be imagine that those keys would be captured by the
> compositor and send only to the applications requesting them. This may
> however break in-focused-application shortcuts.
> Finally, it should be reminded that those keys are in a limited number
> as it generally correspond to special hardware keys
> (play/next/previous/volume control/back-light control/display control)
> that some keyboards provides. The user should never be able to use those
> in a password and it should expect them to work only with special
> applications (media players, screen back light control...).
> I may have overlooked some use cases; please correct me if I'm wrong.
> Tim

Now, lets say i have video and audio player opened, both have no
focus, both register "play" semantic action, and it is same sequence
for them, right?

Now, which of those will get input event if i press combo for "play",
none of them/all/random one, first to register?

I've no idea what would be sane behavior in that case.

Or do you mean, that action is available to just focused application,
and it is just oridinary shortcut, just that is consistent for all
applications, that request it?


More information about the wayland-devel mailing list