Has anyone seen this OS5.0.5/Digi 8e problem

[edfair]

Have multiple Pentium 166 with fully patched 5.0.5 running Digi8e for modem and terminal support under MPI 3.2.3.
All have passthru printing on ttyi1E, the remote access port using a USR 33K modem for dialup.
All 5 machines show wyse60 as the terminal type for passthru printing when port definitions are accessed via MPI from the console but 2 show ansi as the terminal type when accessed via MPI through the modem.
Earlier OS (3.2v4.2) and terminal controls on these same machines were uniform in showing wyse60.
I've downloaded the latest drivers to load asap for test but was curious if anyone else has run across the problem.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[motoslide]

Any chance the TERM variable is being set as they log in?
Check /etc/profile and $HOME/.profile

What about the contents of /etc/ttytype for both ttyi1E and ttyi1e?

We use Digi Xem and C/X.

 

[edfair]

TERM is being set from /etc/ttytype as I log in as root.
`eval tset -xxx`
export TERM

and env gives wy60 which is the ttytype parameter for both

but mpi shows the term variable to be ansi from ttyi1E and wyse60 from the console. But feedthru is going to the screen rather than the printer. All the remotes are wyse.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[motoslide]

First, the disclaimer:
This could be different for the Digi 8e.

On the C/X (and Xem) devices, we have a directory called /etc/digi/epca. In that directory is a file named CX11A.prnt. This is the file which contains the definitions to turn on/off the printer port of the connected device. It also contains the TERM variable related to that purpose.
I've seen this file get corrupt and cause me pain. I believe you can just delete the file before running mpi (or pcu) and it will be regenerated.
This doesn't appear to explain why you see different values depending upon which device you are connected, but it might be worth a shot. Just rename that file in case I'm crazy and the system gets really cranky.

Another interesting test would be to comment out the "eval tset" command from your .profile and see if you get different results. You said in one case the TERM was "wyse60" and in another "wy60". You and I know they are virtually the same, but the Digi utility might be looking at the terminfo entry of each and perhaps one doesn't have the expected flags.

I think you can also set the desired printer on/off codes using the "ditty -onstr" and "ditty -offstr" commands in your login script.

man ditty

Good luck.

 

[edfair]

The wy60 is the termcap entry and the wyse60 is in the printer definition of MPI.

The termianl types do come out the same no matter which way I go in. It is the passthru terminal type that shows up differently. And from what I've seen since first posting it appears to be the first terminal port that has passthru that has the problem, not the remaining 7.

I will try to find the equivalent file on the next system that I bring up in test mode. I'll have one in a day or two where I can do some experimenting without travel involved.

Thanks for the suggestions. I will follow up with whatever solutions I find.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[edfair]

Reload of MPI resolved the issue. Terminal type in MPI for passthru term changed when the user came in over the modem.
I was sitting on the console going in and out of MPI when it did it.

Ed Fair
Give the wrong symptoms, get the wrong solutions.