Re : TestStand 3.5 Runtime object releasing problem

[Atko]

Has anyone ever experienced something similar?

Tools:

Win XP

TestStand 3.5 runtime only (for test deployment)

C# Operator Interface (VS 2005)

Symptoms:

After the operator interface exits, it won't restart, the exe is just sitting in there but doing nothing. I believe it has problems with some property objects with unreleased references, but I can't find a way to figure out.

If the OI runs on a PC with development system installed on it, TS is supposed to pop up the debug warning window with the list of unreleased references and other issues on exit as the "Report Object Leaks" in the Debug options is turned on. It actually can be seen for a "fraction of a second" but it seems impossible to make it stay on the screen after the OI exited, so it appears that the application has nothing to complain about. Also, on the dev PC the OI restarts with no problem, probably because the development system gets rid of the references without letting me know about it

To make it a little more complicated I played a little with the the NI out-the-box C# OIs on the runtime-only PC and loading simple sequences with no Ax refs at all in them, or even without opening any seq files, the OI will not restart after exiting. So it seems it may not be me not properly releasing refs. The GC timer is ticking anyway and triggers the forced GC, so It should not be an issue... Plus: the original NI LV OI behaves the same...

Experience, ideas to share? Anyone?

Thanks,

Atko 

[fungiscience]

Hello,

I got a similar problem with the Debug Warning Window. I enabled the "Report Object Leaks" option. When closing TestStand, I can see the Debug Warning Window appear for a fraction of second, but can't get a hold of what is written in there. I'd like to know what are the leaks observed so I can correct the problem in the code. Any idea what can cause this problem?

I'm using :

• TestStand 4.0.1f1 with customized StationCallback and Models
• LabView 8.5.1
• Proligent 4 RTM10 with Proligent TestStand

Thanks,

Sebastien

[MartyMcDAX]

Hello fungiscience,

This warning often occurs when you use the "Variant To Data" vi to convert a variant to a property object.  This can cause the memory associated with your wire to never be released.  For a more thorough description, you can refer to this KnowledgeBase Article: LabVIEW UserInterface gives PropertyObject Warnings when using UIMessages.

[MartyMcDAX]

Hello Atko,

This may be happening because you are using the .NET example shipped with TestStand 3.5 and intended to be used with the .NET Framework 1.1, but instead using this example with the .NET Framework 2.0.

Fungiscience,

Your object leaks are most likely appearing because you need to close the TestStand Engine reference in your OI prior to closing the OI completely, as described in this KnowledgeBase Article.  If this OI was upgraded from TS 3.5 or an earlier version, then this is probably where the leaks are coming from.  You need to get an Engine reference in your OI and use the Close Reference VI to close the Engine prior to exiting your application.