Impossible Behavior

[Pete]

:headbang:

There's a lot of history to this, but the upshot is that I have a cluster of controls that don't work (anymore) when wired to a certain subVI. A probe on the wire (between the control and the subVI) does NOT register changes to the cluster's control values. However, wiring instead to the loop border, the probe DOES show changes to the cluster's control values. I have never seen anything like this. And it should be impossible! There are no broken wires; the code runs in every other respect, as far as I can determine.

Please, any discussion at all is welcomed.

[Jim Kring]

Please post an example that demonstrates the problem.

[Pete]

Please post an example that demonstrates the problem.

Thanks for the response, Jim. I haven't been able to create a stripped-down version of the code that demonstrates the issue. I think the VI code represents peculiar factors which interact to create the problem. I've run into this kind of situation a few times before--especially in the PDA toolkit (7.1). Everything works great by itself but blows up when you put it all together in the context of other code.

I've been working this issue furiously for the past few days, posting, besides here, at the NI devzone, where a number of participants have weighed in on the discussion, providing much needed perspective (I tend to focus myself into a box when under pressure). You can find it in the LabVIEW forum, topic, "Insane object with a symptom". But I can give you the quick and dirty:

LV6.0.2--medium sized app (~420 VIs) developed years ago, undergoing modifications.

Previously working code snippet in subVI (>2MB size) no longer working.

Insane object errors (2) when trying to save after editing, citing "waveform chart"; deleted chart, no more insanity, but code section not working.

Code section includes a small subVI (~300k) that accepts a cluster of controls (bool, num, str, ring) and outputs same cluster; this output is wired to a local variable of the cluster of controls (to restore button states after user press).

Tried recreating cluster control from scratch using the more modern LV components; tried copy and paste of small subVI BD to fresh VI; tried empty subVI with only cluster control inside to occupy terminal on connector pane--all to no avail.

Deleted local variable--WORKS!

So I set the small subVI up with a strict reference to the cluster control. Now reading and writing to the cluster control is all done internally to the small subVI--no wires connect to the cluster control.

Recreated the waveform chart.

Now all is back to normal.

I can provide the offending VIs, but it could be 3 MB or so. I don't know the size limit if any imposed by this site. If you're really interested in pursuing this wild goose, I'm game.