* Sebastian Bremicker <[EMAIL PROTECTED]> [080222 21:36]:
> I don't know whether it is an X.org or ratpoison problem, but for some
> months now I have serious usability problems with ratpoison keybindings.
> First there was this deadlock, which was an X error, after this still my
> key-combinations with ctrl-alt ceased working after using c-t : or !, now
> there is still one problem left, which seems to be reproducible:
>
> All c-t keybindings stop working when I am in a Gnome or KDE program like
> gnome-terminal or sylpheed, when I quit this program, they work again.
>
> Are there any tests I could run?
Do I understand you correctly that this happens both with 1.4.2 and the
new 1.4.3 you reported against?
(1.4.3 has the keyboard grabbing radically changed so when it happens
with both this is quite an important piece of information).
Besides other newly introduces X errors an other cause for such
behaviour I can imagine is that those programs change the keyboard
mapping somehow (though I guessed this would be done by gnomish and
kdeish window managers instead of the libraries, but I simply not know).
Though even if that is the case, I can hardly imagine C-t being hit by
that. (On second thought, perhaps when very strange things with the
modifiers are done).
Debugging something like that is quite hard. Some ideas:
try to bind another top-level dir, like
definekey top F8 echo catched
(in your .ratpoisonrc or after C-t :)
This should cause it to print a message "catched" in the message window
each time you press F8. It would be intresting if that behviour survives
starting a Gnome or KDE window.
If that works, one could try to change the prefix key from C-t to some
key without modifiers (for example using "escape F9") and testing if
that works (to find out if it are really modifiers, or grabbing the key
while waiting for subsequent keys fails)
Also intresting would be xtrace's output on ratpoison while that happens
and on those programs. Though that would most likely be veeery long, a
bit tricky to start, and questionable if it gives that much information.
(Especially as the more sophisticated xkb extension is not yet
implemented in xtrace, though one should still see what ends up in
ratpoison)
Hochachtungsvoll,
Bernhard R. Link
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
