LabVIEW
Correct Shared Variable behavior
Hi LV_Pro,
you should see it as a good reason to think about the error cluster and what can happen. If you use the error cluster you should everytime think about, can there occur an error or not. If you come to the point there can, then you should think about what should happen then. The Styleguide show you how to handle error inputs. You can also filter some errors to only react on others.
You can't change the behavior now and it's well documented, so it seems to be correct now. Ok, the way to introduce it colud have been better, but we can't change it now and i think it's a good point to change the point of view. 
Do not be cross with me.
Mike
Mike,
I'm quite familiar with LabVIEW style issues, having programmed continuously in LabVIEW since its release in version 2.5 on the Windows platform in 1992. My point is that there are other functions that just pass the error, with it having no effect on the function's behavior, and that is how the Shared Variable did function, for 3 major releases. Whether philosophically this was correct, it was as it was and the documentation did not say otherwise, and to change such a major behavior, with so little "announcement" is what concerns me. I am quite familiar with how the error cluster is used, how to mask specific errors, etc., but one use is to force execution order and what upsets me is the fundemental change, without any apparent "announcement of the change". Had the current behavior been the original desired behavior, and the previous behavior been a bug, and the current behavior been the way it had always been documented, then my using it this way would be an error on my part. But, as mentioned, there are other functions in LabVIEW that just pass through the error, and attempt to execute regardless, so that the original Shared Variable behavior did not seem inconsistent. As the original help documentation didn't specify I actually made a test case, back in 8.2, to see which way it behaved.
