interesting stuff..

From: Mike Benjamin (
Date: Sun Aug 19 2001 - 03:09:54 EDT

I've found a few interesting things while playing with sysmon/sysmond
over the past few days.

When specifying an 'upcolor' on 2 Linux machines the server has a
Segmentation Fault. When specifying under FreeBSD it spits out:
"sysmond in free(): warning: junk pointer, too low to make sense." and
apparently stops reading the config as it complains about the lack of
root indefinitely. Problem is the FREE(upcolor) in loadconfig.c.

When first starting the ncurses client (independant of OS, tested on
FreeBSD/Solaris/Linux) or server (FreeBSD/Linux), the Time Failed of
already occurring alarms will be equivalent to 0/Never/etc.. Refreshing
once will fix it, or if starting the client before the daemon has a
chance to find any alarms it won't occurr.

Refreshing twice quickly under the ncurses client in Linux will cause the
same "zero time" problem to occur, with the same fix as above available.
Refreshing twice quckly under Solaris will cause the client to lock up
completely, requiring a ^C to exit. Refreshing twice quickly under
FreeBSD will cause the Time Failed to change to the current time.

Under FreeBSD/Solaris/Linux the client doesn't leave the term "nice" when it
exits cleanly. You've got to `reset` to clean it up.

One suggestion I have would be a per-user authentication, whether that be
it's own password file, the existing password file, pam interaction,


This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:14:07 EDT