"void wire"

[johnfr]

I have ran into a strange problem with a "void wire" An application that I have had running for months suddenly developed a problem with what it is flagging as a "void wire".  A wire that connects a cluster to a subvi inside of case structure appears broken but the run arrow does not indicate a broken vi. If I turn on help and hover over the wire with the wiring tool the popup says that a void wire is not a broken wire; it is a wire that Labview does not yet know the data type. I have tried deleting and reconnecting the wire, deleting the case structure and re-wiring case structure, nothing helps.  In fact, when I recreated the case structure, then any wire I attempt to re-wire inside the case structure becomes a "void wire".  I have consulted tech support and NI's advanced support guy had not seen this. Curious if anyone out there has seen this?

[Ben]

I have not seen that since about LV 7.0 or earlier. It also manifest itself (if it the same issue!) as "wires from nowhere" where the wire is OK but has no source.

So if it the same, I ended up restoring a backup.

If this is not LV 7.0 or earlier plase share your version number and a sample VI with the wire.

Ben

[johnfr]

Sorry I didn't include version. This is LV 8.6.1.

[Ben]

---

johnfr wrote:
Sorry I didn't include version. This is LV 8.6.1.

---

Well that is intersting!

Could you post up a demo and do you have any idea how to accomplished such a feat from scratch?

If we can get NI to duplicate the issue, the door is open to getting a fix.

Ben

[johnfr]

I have not been able to accomplish it from scratch.  I have tried several times and it hasn't repeated except within the original vi. I have deleted portions of the original vi and tried to re wire. When I do that "everything" I attempt to wire into the case structure becomes a void wire. Because it is part of a proprietary system I can't post the entire program on a public site like this. I have contacted Ni's advanced tech support and they had not heard of this. My tech support guy is currently consulting with ni's R&D. I have sent all the vi's from the project to NI's advanced tech support.

I posted here to see if anyone out in the "real world" had ran across the problem.

[smercurio_fc]

Perhaps it's a corrupted VI? Have you tried to do a force recompile (hold down the Control key and click on the Run button in the toolbar), or to select the entire block diagram and drag it onto a new VI's block diagram?

[Darin.K]

Corruption is also my first guess, but out of curiosity:

Could you post just the cluster constant and the offending case structure, copy and paste the two to a new VI.  Assuming it still shows the problem.

Linked tunnels were new in 8.6 and I can not imagine anything going wrong there, but who knows.

Can you replace the offending case structure with a stacked sequence and then replace again with a case structure.  You will have to fix the cases to be True/False again.

Does it run?

Finally, I find it interesting that the void wire is not the thickness of the usual cluster wire, seems like it should be.  But then again, it seems like it shouldn't be a void wire at all.

[tst]

One way of creating a void wire is by having a subVI with something wired into it and then removing the input from the connector pane of the subVI. This creates a proper wire with a source, but with no sink.

If memory serves, though, it should definitely break the VI.

[JÞB]

Two suggestions:

1) run the VI analyser just to see if there is hidden code under the case structures borders

2) try deleting the tunnels, CTL-B, and recreating the wire.  

Recreated here in 2009

[johnfr]

I agree corruption is the most likely cause. I am attempting to get a clean historical copy from our server.

I tried copying the case structure and cluster constant into a new vi. When I do that the problem goes away and I am unable to duplicate.

I also tried your suggestion of replacing case stucture with stacked sequence. As soon as I did that all of the lines internal to the stacked sequence became void wires. ( A strong indicator of corruption.) When I tried to close LV it stated giving insane object errors for all of the void wires. ( Again a strong indicator of corruption.)