OK, not sure if there is a relation, but it appear on a relatively powerful machine (the one that has most CPUs among those you tested). Unfortunately, I have used one that is pretty similar, too (with no repro).

Regarding the virtual machine: My thinking was that if you could create a VM that shows the problem, it should then behave independently on the host computer - so if you send it to me, I should be able to reproduce the problem as well. Whether you can make this VM simply by creating it on the host computer that has the problem - I am not sure about it, the chances are not big.

Are you willing to install a Debug build of QuickOPC plus Microsoft Debug C++ libs? There is a possibly (but not certainty) that it will show us more details to the problem.

Thank you

Hi,

Thank you for additional info. Are there other differences between the two installations? Specifically, as I suspect some kind of threading-related problem, what is the number of cores and logical processors on the two machines?

What you are seeing should not happen with our component, independent of possible problems with any OPC server. I also do not think there is anything "wrong" with your computer or installations.

Unless we find some other way (what about the VMs?), there is one more think that can be done: I can send you a Debug build of the component. The idea is that it has more "checks" inside the code (ASSERTs etc.), and thus it may come with earlier - and more to the point - error message. I do not want to go this way too early, because it is a bit of a hassle - it also requires an installation of Microsoft Visual C++ Debug libraries on your computer. But it can be done.

Best regards

If I could reproduce it here, I will fix it. Isn't your computer a virtual machine, or is there any chance you can turn into a virtual machine, or create a separate VM which has the same problem?

I have created more tests, and tried various combinations. I could get a problem in the situation when I expected it (two BrowseServers calls), but in other situations. I have fixed this, and I think it is worth trying on your side. The new build is labeled 5.23.1162.1, and it is now the publicly available build in the Download area on your side. (the same fix will go to version 5.30 but somewhat later).

Can you please download it, rebuild your app, and try again. Please be absolutely sure that you are using this new build then - check the full assembly version, both in GAC (it is there by default), and/or wherever you app is located (if the assemblies are copied alongside the EXE). You can right-click on one of the assembly DLLs, select Properties, and then see the version on the Details tab.

Best regards

I have tried in SharpDevelop but still could not reproduce it.
The code change I intend to make is not (at least the way I see it) related to the fact that item subscriptions run in parallel with BrowseServers. It has to do with more BrowseServers calls at the same time. But it is a bit more complicated than that, so maybe.

I will create a program (or modify yours) to actually give it a bigger press (invoke all these operations automatically, without user interaction), and run it then for some time. Hopefully it will show the problem, and if so, will fix it. If not, I will just make the one fix I already know that needs to be made, and make a new build available to you.