ksh acting weird under SCO Unix

[Sean Fagan]

In article <35582 at cup.portal.com> ts at cup.portal.com (Tim W Smith) writes:
>We obtained the V.3.2 release 2 update recently (I think that's
>the one.  Whatever is the latest).  I decided not to install it,
>however, because the development system update is backordered.
>I extracted ksh from one of the installation floppies, however.

Uhm, why do you think this will work, at all?  In general, I would be leery
of installing parts of an OS onto another OS.

In this case, it really won't work (as you note).  3.2v2 adds working job
control, and ksh was built to use it (it works nicely, btw 8-)).  In
particular, ksh now will have every "command" be in a seperate process
group, so that job control works properly, if the '-m' option is set (set by
default in an interactive shell).

3.2.0 and 3.2.1 had some serious bugs in relation to job control; as a
result, none of our utilities used it.  3.2v2, however, has those bugs
fixed, so things do use it.

ksh exits, as you describe, because it no longer has access to the
controlling tty.  This is because of a bug in the code to change the pgrp of
tty's (fixed, obviously, in 3.2v2).  The bug causes, basicly, ksh to be
unable to get the tty back after it's run the process.  Because it can't get
to the tty, it dies (not necessarily gracefully, obviously 8-)).

Most of the problems you'll run into are related to that.

If you *must* use ksh, do a 'set +m' immediately when starting the shell
(either as the first command at the prompt, or in your $ENV file), or
upgrade to 3.2v2 (8-)).

-- 
-----------------+
Sean Eric Fagan  | "*Never* knock on Death's door:  ring the bell and 
seanf at sco.COM    |   run away!  Death hates that!"
uunet!sco!seanf  |     -- Dr. Mike Stratford (Matt Frewer, "Doctor, Doctor")
(408) 458-1422   | Any opinions expressed are my own, not my employers'.