Re: Annoying mouse binding problem

[Harald Hanche-Olsen <hanche math ntnu no>]

+ Timo Korvola <Timo Korvola iki fi>:

> Harald Hanche-Olsen <hanche math ntnu no> writes:
> > What seems to happen is this: Starting at some time (typically after
> > unlocking xlock in the morning), sawfish can no longer see button
> > presses arriving in a window.
> 
> In any window or some particular window?

Every window is affected, including new ones created after the problem
crops up.

> Basic window configuration such as location, size and stacking order
> is maintained on the X server.  Additional configuration such as
> workspace numbers can be stored on the server in X window properties.
> You can have a look with xprop(1).

Thanks. I'll poke around when I have more time, which unfortunately is
not now.

> You might try checking (window-get <affected window> 'keymap) and the
> global variable window-keymap in sawfish-client.  If the window keymap
> looks OK, there might be some problem with grabs.

They do look OK, to my untrained eye. For example,

(let ((km (window-get (get-window-by-id #x160000d) 'keymap)))
  (mapcar (lambda (e) (list (car e) (cdr e) (event-name (cdr e))))
	  (cdr km)))

returns

((move-window-interactively (3 . 67240192) "W-Button1-Move")
 (move-window-interactively (1 . 139264) "Button6-Click")
 (move-window-interactively (109 . 33619968) "Super-m")
 (resize-window-interactively (114 . 33619968) "Super-r")
 (resize-window-interactively (3 . 67240193) "W-Button1-S-Move")
 (toggle-window-shaded (65362 . 67174401) "W-S-Up")
 (resize-window-interactively (1 . 139265) "Button6-S-Click")
 (delete-window-safely (65535 . 67174400) "W-DEL")
 (send-to-next-workspace (65363 . 67174404) "W-C-Right")
 (send-to-previous-workspace (65361 . 67174404) "W-C-Left")
 (popup-window-menu (1 . 67240960) "W-Button3-Click")
 (lower-window (65364 . 67174400) "W-Down")
 (raise-window (65362 . 67174400) "W-Up"))

None of the four button bindings listed work. (Initially they do, but
after a while they stop working.)

> Unfortunately grabs are rather difficult to debug.

I can imagine. But surely, since the difficulty remains after a
restart, the problem must be in data stored on the X server?
Unless it's a bug in the server itself, of course. As I said, I'll
poke around and see if I can find anything suspicious. Thanks for your
help this far.

- Harald