Skip to main content
Question

AUTOLOGOUT behavior

  • November 24, 2025
  • 4 replies
  • 28 views
Steven Douglas

Does anyone have an accurate handle on how and when AUTOLOGOUT triggers?   The manual says….

Use AUTOLOGOUT to enable or disable automatic logout. With AUTOLOGOUT enabled, UniVerse logs you out automatically if you have not pressed a key within a specified time.

But I don’t believe that is accurate.   We’ve had some confusing results in testing.   A process on a loop with periodic TIMEDATE() displays (output only) does not trigger.  Similarly a loop process with no screen output but just a SLEEP X also does not trigger even in between loops before it goes back to sleep.

So, we are trying to understand the actual nuts and bolts of how/when it triggers.   Thanks for any help.
Steve Douglas
VP Systems Admin
MBMS, LLC

4 replies

John Jenkins
Forum|alt.badge.img+1
  • John Jenkins
  • Participating Frequently
  • November 25, 2025

Syteve, 

 

I don’t know whether his helps, but from the 14.1.1 release notices if you use NLS:

 

UNV-34431 - Windows only. Beginning at 11.3.4, the AUTOLOGOUT function can fail to execute when NLS is installed on the system. When the AUTOLOGOUT timer expires, a control character is display and the session incorrectly remains active. As noted above, setting the AutologoutInputFlag registry key to 0 can be used to work around the issue at this release. This issue is being investigated to address this behavior in a future release.

Regards

JJ

Steven Douglas
  • Steven Douglas
  • Author
  • New Participant
  • November 25, 2025

Thanks for the response John,   I don’t think that applies here.  I should have put in the original post, we are running Universe on top of Redhat.
With further testing, I am starting to suspect that the timer does not even start to run unless/until the process is at an input prompt.   That would kind of make sense based on the description in the manual that it is keystroke based.   It definitely triggers correctly while sitting at TCL and at an INPUT statement in a basic program.   I’m still testing to see what might happen in an INPUTIF scenario etc….     So, if my theory is correct, AUTOLOGOUT would never trigger on a process that is actually doing something for a long period of time but is not at an input prompt.

John Jenkins
Forum|alt.badge.img+1
  • John Jenkins
  • Participating Frequently
  • November 25, 2025

Steve,

That would make sense I think - as you would not want to log out a processing intensive process that updates multiple associated files that is running for a good while just because there have not been any key depressions. To do so could well leave files in an inconsistent state.

 

I’m unsure what an input buffer poll (e.g. INPUT var,-1 and SYSTEM(14)) might do to the autologout timing - I’ll have to go and have a play, and maybe a play with the edge conditions of what happens if the clocks go forward or backward due ti daylight-saving while AUTOOGOUT is running and sitting at TCL (because my  mind thinks like that).

Regards

JJ

Steven Douglas
  • Steven Douglas
  • Author
  • New Participant
  • November 25, 2025

John,
Thanks for the thoughtful response.   And thank you (I think) for the daylight savings and SYSTEM(14) thoughts.   I hadn’t thought of those.
For further info.   INPUTIF has no impact and nothing appears to trigger when passing an INPUTIF which also makes sense.   My mind is also now questioning what might happen if input was required in a linux “shelled” command within a basic program or a PROC or something.   Or is it just the straight INPUT or TCL that starts the timer?

Terms & ConditionsAccessibility statement