Strange Behavior Checking Running State of VI

[GregFreeman]

I am working with some legacy code that I didn't write and am seeing a strange issue but really am not sure where it has come from. In the last year I have worked with this code, this hasn't come up and I haven't changed anything that I think would effect it.

In this code, VIs are launched dynamically from a launcher loop using the Run VI method. When the dynamic VI is finished, it sends an event back to the launcher loop, which is then responsible for dynamically launching the next VI. When the launcher loop receives the message that the dynamically launched VI is finished, it polls that VI's Running state until that VI's state changes from Top Level VI to Idle before launching the next VI.  It does this polling with a two second timeout.

This all works fine most of the time, but then randomly the state of the launched VI will continue to be returned as Top Level VI, not Idle, even though from what I can tell the VI has stopped. I proved this by logging a timestamp to a log file as the very last thing my dynamically launched VI does. I then logged a timestamp after the timeout and they are 2 seconds apart. So this tells me even though I think the launched VI is finished, its Running state as seen by LabVIEW is not changing to Idle.

I am aware this is not the best way for this to be implemented, but as I mentioned this is a legacy system and I really don't have the ability to refactor the entire application. I'm just trying to grasp an understanding of what may be going on. For now, I've just increased the timeout, hoping this will mitigate the issue in the random cases that this pops up. But I'd really like to find the root cause.

[drjdpowell]

Do you use any DVRs?  I believe quite a lot happens when the LabVIEW execution system changes a VI from "running top level" to "idle", including destroying any references created by that VI.  Not sure what happens if a DVR that it owns is currently being accessed by another VI; I assume it must wait for the DVR to be released before finishing teh transition to "idle".

[GregFreeman]

I don't believe there are any DVRs, but I can't be certain. There are however various references types everywhere. That being said, for now I have just extended the timeout and it seems to be fine. I am going to assume there is something going on akin to what you mentioned in some way, shape, or form. The code base is a tangled mess and trying to get to root cause may just be a futile exercise.