Re: why does M-w (kill-ring-save) hurt so bad?
On Tue, 2002-12-17 at 19:30, Georg Nikodym wrote:
In case anybody cares, the thing that drove me to the dark side:
objdump -d vmlinux > /tmp/vmlinux.S
load /tmp/vmlinux.S into XEmacs (with font-lock, lazy-lock/shot et al.)
Mine takes about 15 minutes and bloats to 140MB (this is a 1GHz laptop
with 512MB RAM).
Huh? FWIW, with similar settings (font-lock, lazy-shot), on a
P4/1.85GHz, 512MB RAM, XEmacs 21.4.11rc1 gives me:
$ time /usr/bin/xemacs /tmp/vmlinux.S -eval '(kill-emacs 0)'
real 0m3.377s
user 0m2.512s
sys 0m0.172s
...and GNU Emacs 21.2.1:
$ time /usr/bin/emacs /tmp/vmlinux.S -eval '(kill-emacs 0)'
real 0m1.819s
user 0m0.805s
sys 0m0.076s
The memory usage values (SIZE from top) are about 45M for XEmacs and 30M
for GNU Emacs with this file loaded. And /tmp/vmlinux.S is about 23M.
With XEmacs 21.1.14 the load time is about 2.1 seconds. Yes, there's a
noticeable difference, but nowhere near 15 minutes and 140M...
--
\/ille Skyttä
scop at xemacs.org