phenomenological behaviour of labview

[agony]

to be or not to be...

as attachment some window-dumps from

1. a very simple VI, that get's an object in, unbundles it to get a certain attribut and the exits.
2. a probe from the object near the input
3. a probe from the unbundled attribut
4. a probe from the object near the output


as you can see, the attribut "_valid", a boolean value, is "true" in (2) and (4), but the unbundled one (3) is false.

well maybe I should discuss some philosphy-topics with LAbVIEW instead of trying to program with it...

P.S.: because this forum doesn't allow more than 3 attachments, you have to do it without 4th window-dump, but you can believe, that it's the same as 2nd one.

[Sean]

My guess is that you've used an unwired bundle function.  This means you've replaced its value with the default value - false for a boolean.

Good word though - Phenom...

[ssk]

Could you post the vi?

[TCPlomp]

---

@Sean wrote:

My guess is that you've used an unwired bundle function.  This means you've replaced its value with the default value - false for a boolean.

Good word though - Phenom...

---

That is not true
Quote from the 'Bundle' help:
"When you wire an existing cluster to the middle terminal of this function, the inputs are optional. LabVIEW replaces only those cluster elements that you wire."
Although I'm not sure about the behaviour with classes.

Ton

[Sean]

Fair enough - I stand corrected.  I tend to stick to the Bundle By Name function myself.

I've no knowledge of this object though - does it stick to its rules in this case?  Would the Bundle By Name function work with it?  Might be worth a try.

[agony]

I could post the VI itself but it won't run on its own, because it's part of class-framework, I currently try to build up.

but...
meanwhile I've made changes to the VI, especially to the unbundling-function and now it seems to work.

maybe I've got a backup in an previous state, but I think this "behaviour" of LabVIEW is not completly deterministic.

for example I've got the "feature" last time that after some other changes the property-nodes of some elements shows the right referenced element, but LabVIEW believe, that the conenctions are bad. I had to change the reference first to another element and then back, then no error was shown any more...

so the picture from the VI above should show the necessary things...

such things as the malfunctioning unbundle are really nasty, because they're diffcult to identify and they cost much time, that I don't have anymore...

[Sean]

I'd be fairly confident that LabVIEW isn't behaving randomly.  Zip it up if you like.