On Thu, 30 Oct 2003 20:17:05 +0900
"Stephen J. Turnbull" <stephen(a)xemacs.org> wrote:
Please keep the list as a CC, at least. I know about a fair amount
of
XEmacs, but for everything I know, there are other developers and
users who know a lot more about it. "Many eyes, shallow bugs."
My mistake, I didn't notice the cc.
>>>>> "junq" == junq <junq(a)ihubbell.com> writes:
junq> I'm certainly no authority on mem leaks but this looked too
junq> consistent to me. Anyway I stumbled across this measuring
junq> processes and this one happened to have caught my eye as
junq> unusual.
Well, the problem is that there are memory leaks, and there are stupid
Lisp programs that use unbounded amounts of memory. One thing that
makes me curious is the CPU usage which looks to be about 2-3%. On my
machines (450MHz Pentia) a quiescent XEmacs uses more like 0.2%. I
wonder if there might not be a subprocess or itimer doing evil things.
Name Value Restart Function Idle Arguments
---- ----- ------- -------- ---- --------
"auto-save" 85.31 240 auto-save-itimer no NONE
"itimer-0" 5 5 cperl-get-help-defe yes nil
junq> This has 10 buffers open with small perl scripts, a dir and
junq> some scheme code. I use early a.m. and late p.m. now and
junq> again.
Use C-u M-x (buffer-list) RET to insert the list of all live buffers
C-u M-x buffer-menu RET gave me Invalid function: (4)
in the current buffer, and see if any of them look weird. (Use M-%
SPC #<buf RET C-q RET #<buf RET ! to reformat the list sanely.)
Try M-x list-processes to see if there are any subprocesses that might
be generating junk or using space somehow.
Proc Status Buffer Tty Command
---- ------ ------ --- -------
junq> Here's the latest png FWIW. It's about 5 1/2 days worth.
Ah, much more convenient. Thanks for going to the trouble.
junq> I'd be willing but finding the time will be difficult. Is
junq> there anything I can do with the xemacs I have that might
junq> help getting to the bottom?
See above. I'd really like the output of M-x show-memory-usage to
compare to your VSZ, but that function is only available if you
configure with --debug.
Maybe on the weekend I can find a few spare cycles.
junq> I'll try to help if I can. Just not sure I want to get into
junq> compiling the software. (I'm a little afraid to ask) What
junq> dev. env. is required to build it?
1. Usual GCC toolchain.
2. -devel RPMs for all libraries used (X, Motif, GIF, XPM, TIFF,
JPEG, PNG, ncurses, Berkeley DB, LDAP, PostgreSQL, Canna, Wnn).
You can omit the databases, Canna, and Wnn if you don't use them.
You can omit GIF, TIFF, and JPEG, but they're really nice to
have. Don't omit XPM, PNG, or ncurses, they require a certain
amount of expertise to work around likely problems.
3. XEmacs sources.
My guess is that if you get the source RPM from Red Hat, it will
complain if you don't have what it wants. I don't know how to change
the configuration that they've set up, though.
I might be able to compile and install into /usr/local; if time's
on my side this weekend.
--
Institute of Policy and Planning Sciences
http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.