click operator interface close

[dseel]

Hi:

We are using a modified version of the full featured CVI Teststand interface (CVI 9.0.1, Teststand 4.0.1f1). We are having sporadic problems where the operator interface.exe will crash (Operator Interface.exe has encountered a problem and needs to close). This has been happening for years and we have never been able to isolate the problem. Here is what we know:

* no memory leaks are detected in our associated DLLs or the modified Teststand interface (using CVI 9 memory tracker)

* crashes still occur using the default full featured CVI Teststand interface

* operators report seeing crashes daily

The key piece of new information is that we now have regression stations running automatically 24/7 in our labs. These stations run for days on end without any problems. But if we are logged into the station and click on the operator interface or one of its associated windows, at this point we sometimes get a crash.

So is it possible that a mouse event, or enabling/moving an assoicated panel with the mouse, could cause a crash like this? If so, how should I go about isolating this problem so I can fix it?

Thanks,
Dave

[muks]

See whether this discussion will give you any insight

[RayFarmer]

Hi,

So this must have been happening with previous versions of CVI (as CVI 9.0.1 hasn't been around for years).

Have you always been using TS4.0.1f1?

Do you have a bit more information on the meaning of the term crash:

1. Does the UI freeze?

2. Does the UI close without warning and no error messages?

3. Are there any warning / error messages.

Is there any test sequences running at the time the crash occurs?

Regards

Ray Farmer

[dseel]

Hi:

This was definitely happening before CVI 9 and before TS 4.0.1f1. Crashes have always existed, but it is very hard to track how often they occur, so I don't know if the number of occurrences has gone up or down the latest versions. 

I've attached a screenshot of what I mean by a typical crash. We do have cases where the UI freezes, but the screenshot is more common. This screenshot was taken when we were trying out the default Teststand CVI full featured interface to see if the crashes were being caused by one of our UI modifications. I guess not. 

There most definitely are test sequences running at the time the crash occurs. But the crash occurs on a wide range of test sequences, so I don't think the crash is caused by a specific sequence file. We have made modifications to the process models and associated DLLs though, so it is possible a common library is the problem.

muks,

As far as the other discussion goes, it may or may not be relevant. We are not using labview at all, but we definitely are using separate threads (from both CVI and Teststand) quite often. And these threads frequently access the Teststand API in some way (grabbing sequence variables, etc). Some of our libraries use thread pools in CVI to run active x applications (windows media player). Anything in particular I can do to isolate if multi threading is the culprit here?

Thanks,

Dave