Re: Causing (and detecting) a workstation failure

Joe,

I think the problem is that the code never advances past the EXFMT.

You can't trap for crashed session even if it happens between the time you
press enter (or another key) and the time you test for in90, which is
microseconds.

if the session crashes when the code is sitting at EXFMT waiting for user
action, it continues to wait, since the user didn't hit a key.

I do remember something about a shutdown indicator (or status) that you can
test for or use in a special error handling subroutine?  this might be more
along the lines of what you are looking for.

on the other hand, I've just always looked for a different way to recover
lost sessions.  code for the eventuality and develop recovery proceedures
to use when needed.

hth

rick
---------------------------------


This should be relatively easy.  I'm trying to test for when a user shuts
off a workstation.  I've coded an error indicator in the EXFMT, and am
testing fo the indicator.  If an error occurs, I send a message to *SYSOPR.

Unfortunately, I never get there.  As soon as I shut off the tube (that is,
as soon as I disconnect the CAX session), the job ends and I don't seem to
get to my *IN90 test.

How quickly does the system kill jobs when they lose connection?  Shouldn't
there be an opportunity for cleanup?  Or am I coding incorrectly?

Relevant snippets:

fattrd     cf   e             workstn infds(Infds)

D INFDS           DS
D XSsts             *status