Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!tut.cis.ohio-state.edu!snorkelwacker!spdcc!merk!alliant!linus!cvbnet!caribe!aperez
From: aperez@caribe.uucp (Arturo Perez x6739)
Newsgroups: comp.unix.wizards
Subject: Tracking processes started by user OR batch job?
Message-ID: <291@cvbnet.Prime.COM>
Date: 21 Dec 89 20:08:47 GMT
Sender: postnews@cvbnet.Prime.COM
Lines: 32

We are trying to determine, programmatically, the "relatedness"
of several processes under Unix.  The scenario is as follows:

	% <user enters command to check into data processing system>
	% <user executes data processing command #1>
	% 	.
	%	.
	%	.
	% <user executes data processing command #n>
	% <user enters command to check out of data processing system>

Now, what we're trying to determine is for data-processing-command k,
1 < k <= n + 1 (the '+ 1' is for the signout), whether k is

	1) started by the same interactive process as was the
		"check into" process, i.e. top-level shell.
	2) If all of the commands were executed in "batch" (i.e. non-
		interactive), what's the common underlying thread? 

That is, was k started by the same "interactive command processor" (as the
manual says) as the check-in process.

This gets complicated under the CSH because for commands executed directly
by the interactive csh because CSH does different things when it's interactive
vs. non-interactive.

Any ideas out there?

Arturo Perez
ComputerVision, a division of Prime
aperez@cvbnet.prime.com
Too much information, like a bullet through my brain -- The Police