Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!mit-eddie!bloom-beacon!eru!hagbard!sunic!mcsun!hp4nl!ruuinf!accucx!pluto.phil.ruu.nl!jeroen
From: jeroen@pluto.phil.ruu.nl (Jeroen Scheerder)
Newsgroups: comp.sys.atari.st
Subject: SYSMON bug with PoolFix4
Message-ID: <782@accucx.cc.ruu.nl>
Date: 5 Oct 90 10:03:47 GMT
Sender: news@accucx.cc.ruu.nl
Reply-To: jeroen@pluto.phil.ruu.nl (Jeroen Scheerder)
Organization: Logic section, Department of Philosophy, Utrecht
Lines: 30

I've just discovered a bug in SysMon, which may be of a general nature,
that occurred when I tried to find PoolFix in my program list.

The trouble was: poolfix was in the table as " rul_Prg", but it really
should be "_OOL_Prg", were it not that SysMon knows about "_OOL" and
tells you it's PoolFix if it finds it.

Now why did SysMon think the XBRA-id of poolfix was " rul"?
Examining poolfix4 with a disk monitor, I found out it contained
some message ".. not following XBRA rules", of which the part
"XBRA rul" looks - yes! - just like a XBRA id (but it isn't).
So I guess SysMon just finds out which programs are hooked into
which vectors, and then searches the program's code for an XBRA
struct, thus falsely recognizing "XBRA rul", since it occurs
in Poolfix4 before the "XBRA_OOL" *REAL* XBRA id.

The fix I made was convert "XBRA rul" to "xbra rul"; I rebooted
the machine, and - voila - SysMon told me about a program
PoolFix being resident.

This problem may occur with other programs containing messages
with "XBRA" as a substring, if it occurs before the real XBRA
id; so SysMon really should check whether the id it thinks is
an XBRA id really is located 12 bytes before a routine which
adress is stored in some system vector.

I post this here, because I did not find a true email adress
of mr. Ivanovic (sorry about misspelling the name) in the SysMon
manual; for now we can all correct the bug in an ad hoc way, until 
it is fixed in SysMon itself.