why does M-w (kill-ring-save) hurt so bad?

Running "xemacs -q" (21.4p8) via remote ssh, over a pair of T1s. Every time I type "M-w" to copy something to the kill ring, xemacs stalls for 1 - 3 seconds. No kidding. I tried (setq debug-on-quit t) and quickly typing M-w C-g and all I get is: Signaling: (quit) kill-new("16-Nov-2002 (sat); 11:36 PM\n") copy-region-as-kill(13756 13784) kill-ring-save(13756 13784) call-interactively(kill-ring-save) Of course that shows up 1 - 3 seconds later; so I don't know if that means that kill-new has gone insane, or if the slow part happens before there, but non-interruptably. Right now I'm seeing C-k and M-DEL exhibiting similar lossage, but I don't think they always do -- or perhaps they always do, and I'm misremembering. At first this made me suspect that "copy the emacs kill to the X clipboard" was on, but it doesn't appear to be. But this does smell like "kill causes excessive X traffic." I have not run an X sniffer, mostly because I've forgotten how. xemacs-21.4.8-16 RPM, Red Hat 8.0. I can't remember if it's their RPM, or yours; perhaps this is a clue: uname -a: Linux astest 2.4.9-e.8smp #1 SMP Fri Jul 19 15:38:30 EDT 2002 i686 unknown unknown GNU/Linux Source code location: /usr/src/build/141167-i386/BUILD/xemacs-21.4.8 I've seen this behavior on two different hosts (depending on which desk I'm sitting at.) -- Jamie Zawinski jwz(a)jwz.org http://www.jwz.org/ jwz(a)dnalounge.com http://www.dnalounge.com/