[patch at end]
[Sorry for replying so late, I have been in culinary heaven, i.e.
Japan. Since XEmacs 21.0 is not out of the door yet I believe it can
and should be fixed]
Martin Buchholz <martin(a)xemacs.org> writes 3 weeks ago:
J> 2. XEmacs Considers a key a modifier if ANY of its Keysyms is
a
J> modifier.
J> [this] is wrong (or at the very least
J> inconsistent with the rest of the X world).
I'm not convinced that XEmacs is acting incorrectly.
Personally is I do not really think it is relevant whether or not
XEmacs is incorrect. XEmacs is behaving differently from
xterm/XIM/other apps which is what matters. They all accept keysyms
even if another keysym on the generating key is a modifier.
113 0xff7e (Mode_switch) 0xff20 (Multi_key)
What is Shift-AltGr supposed to give with the above combination?
Dunno, isn't that something for X to worry about, not XEmacs?
Basically. I think we should do it something like this (I moved the
behavior to the x_to_emacs_keysym function since it doesn't require
duplication of the XlookupString calls).
1998-07-16 Jan Vroonhof <vroonhof(a)math.ethz.ch>
* event-Xt.c (x_to_emacs_keysym): Return nil for modifier keysyms.
(x_event_to_emacs_event): Let x_to_emacs_keysym check for modifier
keys thus no longer considering all keysyms on a key.
Index: event-Xt.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs-20/src/event-Xt.c,v
retrieving revision 1.41
diff -u -u -r1.41 event-Xt.c
--- event-Xt.c 1998/06/20 00:58:46 1.41
+++ event-Xt.c 1998/07/17 03:09:56
@@ -769,7 +769,7 @@
/* simple_p means don't try too hard (ASCII only) */
{
KeySym keysym = 0;
-
+
#ifdef HAVE_XIM
int len;
char buffer[64];
@@ -794,7 +794,8 @@
than passing in 0) to avoid crashes on German IRIX */
char dummy[256];
XLookupString (event, dummy, 200, &keysym, 0);
- return x_keysym_to_emacs_keysym (keysym, simple_p);
+ return (IsModifierKey (keysym) || keysym == XK_Mode_switch )
+ ? Qnil : x_keysym_to_emacs_keysym (keysym, simple_p);
}
#endif /* ! XIM_MOTIF */
@@ -843,7 +844,8 @@
{
case XLookupKeySym:
case XLookupBoth:
- return x_keysym_to_emacs_keysym (keysym, simple_p);
+ return (IsModifierKey (keysym) || keysym == XK_Mode_switch )
+ ? Qnil : x_keysym_to_emacs_keysym (keysym, simple_p);
case XLookupChars:
{
@@ -983,20 +985,16 @@
{
Lisp_Object keysym;
XKeyEvent *ev = &x_event->xkey;
- KeyCode keycode = ev->keycode;
-
- if (x_key_is_modifier_p (keycode, d)) /* it's a modifier key */
- return 0;
-
/* This used to compute the frame from the given X window and
store it here, but we really don't care about the frame. */
emacs_event->channel = DEVICE_CONSOLE (d);
keysym = x_to_emacs_keysym (&x_event->xkey, 0);
- /* If the emacs keysym is nil, then that means that the
- X keysym was NoSymbol, which probably means that
- we're in the midst of reading a Multi_key sequence,
- or a "dead" key prefix, or XIM input. Ignore it. */
+ /* If the emacs keysym is nil, then that means that the X
+ keysym was either a Modifier or NoSymbol, which
+ probably means that we're in the midst of reading a
+ Multi_key sequence, or a "dead" key prefix, or XIM
+ input. Ignore it. */
if (NILP (keysym))
return 0;