Re: key release event "leaked"

[dmb]

Hi,

Thanks for the report!

I think we'll have to think about tthis a bit.  My gut reaction is that we
should slurp both events, but there may be something obscure in the standards
for window managers that we have to respect (I kind of doubt it though).

Does this specifically break your workflow somehow?  If its an interesting bit
of trivia it would be lower priority.  I think it shouldn't be too hard to slurp
the events, but I don't have a whole lot of time to look into it.

Sincerely,

David


cage <cage-dev@twistfold.it> writes:

> Hi!
>
> I do not  know if the behaviour is  a bug or not so I  am sending this
> message here.
>
> When  the screen  is split  (vertically or  horizontally, it  does not
> matter), in my case  using 'C-p 3', the event "the  key for number '3'
> has been pressed" is captured by the WM but the event that immediately
> follows: "the key  for number '3' has been released",  is not captured
> and "leaks" to  the widget that has the focus  in the application that
> had the focus.
>
> To  better  illustrate  this  behaviour  I  have  written  this  brief
> function (TK interpreter (wish) is needed):
>
> -----------✀-----------✀-----------✀-----------
>
> (ql:quickload "nodgui")
>
> (defun main ()
>   (nodgui:with-nodgui ()
>     (let ((entry  (make-instance 'nodgui:entry)))
>       (nodgui:bind entry
>                 "<KeyRelease>"
>                 (lambda (e) (format t "keyrelease ~s~%" e)))
>       (nodgui:bind entry
>                 "<KeyPress>"
>                 (lambda (e) (format t "keypress ~s~%" e)))
>       (nodgui:grid entry  0 0)
>       (nodgui:focus entry))))
>
> -----------✀-----------✀-----------✀-----------
>
> after loading this code and calling the function 'main' a simple
> window with a text entry is shown, the entry get the focus.
>
> Splitting the screen will fire the event "<KeyRelease>" (I do not know
> the way this  event is called in  the X11 protocol, so I  am using the
> TCL/TK term) and  the event is printed on  *standard-output*, you will
> see that  the event has  a char slot  that has a  value of '3'  or '2'
> depending of the splitting command.
>
> I am not  sure that this could  be considered a bug, that  is, the key
> release should  be swallowed by  stumpwm together with  the "keypress"
> event, or it is  up to the GUI application to  deal with this spurious
> "keyrelease" events.
>
> Looking forward for your opinion!
>
> Bye!
> C.