First, since I have no access to your system, Sam, and don't have a
Fedora system myself, I can't reproduce. If you would provide an exact
recipe, starting with 'xemacs -vanilla' and providing the necessary
keystrokes to verify that Something Bad has happened, I'll ask a Fedora
user to verify and help localize the problem.
Second, although it might be an XEmacs bug, it is highly unlikely to
have "come into XEmacs in Fedora Core 10" (unless it's a buggy Fedora
patch, in which case you'll have to talk to them about it). AFAIK,
nothing has changed in XEmacs keyboard handling in a decade except for
certain changes in handling Cyrillic keysyms. I don't know of any code
in XEmacs that would remap the keyboard, again AFAIK XEmacs is
completely passive with respect to the characteristics of the keyboard
and screens. So any change in behavior was occasioned by a change in X
or OS keyboard handling. Recent versions of the X Input Method
implementation are known to be even buggier than the historical
versions; configuring XEmacs with the --with-xim=no flag is strongly
recommended anyway (unless you like random uninterruptible hangs). XIM
is the only facility that XEmacs uses that is likely to affect keymaps,
so it might make things better for you. If so, please let us know.
I suspect that deadkey support is handled by XIM. Try rebuilding XEmacs
using --with-xim. However, you risk inflooping in XIM deep inside Xt;
this is uninterruptible (kill -9 is required). AFAICT that is an Xorg
bug in the Xt event loop (which is acknowledged to be broken, see for
example http://bugs.freedesktop.org/show_bug.cgi?id=20048; I don't think
this has been fixed, just worked around).
Patches to support XKB would be welcome (which *might* fix the random
keyboard remapping problem too, but I don't think it would help with
deadkeys), but it will be a while before I personally can get to
producing them. Failing that, pointers to XKB documentation etc would
be a big help.
|