Charles Waldman writes:
Stephen J. Turnbull wrote:
> AFAIK, the only thing that can have that effect that is under control
> of a client is a call to XGrabKeyboard. There's exactly one such call
> in XEmacs 21.5 I could find, in x-grab-keyboard. There are no calls
> to x-grab-keyboard in XEmacs sources that I found.
Stephen, I think you may have missed something...
Indeed I did, and it's my turn for a "why didn't *I* think of that"
(defun passwd-grab-keyboard ()
In X11, reading passwords is basically *the* sanctioned reason for a
grab. (Windows is much more likely to grab the keyboard or even the
whole console; different UI philosophy.) It's in a package, which is
why my grep didn't find it. But I should have known that it was used
at least once somewhere in XEmacs, precisely for this!
However, read-passwd does guarantee that passwd-ungrab-keyboard will
be called (with an unwind-protect). So I don't really understand why
the grab would linger, unless GTK somehow as a different way of
handling keyboard grabs.
XEmacs-Beta mailing list