Subj : Problems with events running
To : Sean Dennis
From : Mike Luther
Date : Mon Feb 14 2011 10:30 am
Sean ..
SD> Hello, All.
SD> I'm having problems with a single nightly event
SD> running at the wrong (very wrong!) time. I have but a
SD> single EVENTS00.BBS file for the entire system which
SD> is one dialup node and three telnet. The line in
SD> question:
SD> Event All 0200 exit=105
SD> For some reason, this event is running at varied times
SD> between 0800-0900. Do I have something wrong in the
SD> events file?
I assume that you are running this whole deal under a .CMD file which also uses
the numeric value with a program exit to decide where in the .CMD file the
logic should next go to perform the desired action.
OK .. What you can do in many cases to figure out what might be going wrong and
how programs can somehow interfere with a 'proper' logic step operation this
way is somewhat simple!
You create an additional step by step mini-program action at any given generic
point like this to write a discrete little file on your hard drive that has a
unique text message in it. Which you can then use to trace a running .CMD or
REXX file to see where a given error is being created! You give each successive
step a different pointer number, 1, 2, 3, 4 ..66 or whatever. You then use the
little program to write the message you want and that discrete unique pointer
to a common trace file on your hard disk.
Then, sort of in your case, if, for example, the program doesn't exit with
error level 105, as expected, but somehow is different, you can then spot exact
where to go looking for a fix.
This technique can actually be used to manually trace the whole OS/2 operating
system matrix for even a massive multi program mix. Which can then tell you
exactly where a given OS/2 driver or program mix may have botched even the ring
level operations that leads to a strange failure like this.
This is exactly the reason and the technique that was used for the decades I
was on the official OS/2 Development Testcase crew. So that I could check all
the latest driver and op system changes on my multiple OS/2 systems. Including
the full multiple BBS node number mess that was and still is on the OS/2 box on
which this is being written! Which is how we found out a whole huge number of
program development errors for the IBM team going all the way back to where it
was in Boca Raton, Florida!
