Data Value Reference bug?

[X.]

I can't pretend that I am an expert of Data Value Reference (DVR) but here is an intriguing behavior...

Here is an excerpt of my diagram (to focus on the region of interest. I honestly believe the rest of it is irrelevant but will post the VI if asked):

I have basically a large data structure which I want to upload to a processing VI only once. This is what the case structure is for: first call, I make a copy of the data and store it in a shift register of the external while loop; any call after that, I do nothing with the DVR (the DVR wire is basically propagated untouched to a DVR indicator which is handled in the calling VI).

Totally unrelated is the two-pronged Boolean shift register to the right. At this stage, one of the register is uninitialized and my VI is broken... But wait! Take a look at the error window:

What happened to my DVR? Unsupported type???

Nevermind, let me first initialize the shift register and I'll look into that later:

OK, so now my shift register is fixed...and so is the DVR as well!

If I remove the shift register initialization, the DVR error pops up again.

Odd...

[X.]

It appears that these unrelated DVR errors appear in many cases. I just used one in another subVI and while I was drawing up the diagram, I had a bunch of expected error PLUS the same "invalid contained type" error, although the DVR was untouched...

[X.]

Since I am the only one to be entertained, here is a follow up on this story. The diagram below uses the same DVR as above:

Really nothing fancy. But there is an error. Can you find it? If not, you are not helped by the Error List window which shows:

Again, this bizarre "Invalid contained type" error... Notice that it is an FP error... !?

If I finish the diagram and connect the loop index to the Array index, all error disappear.

So, once again, why do I get this DVR-related error when there is absolutely no apparent reason for it?

[Ben]

Those are curious X.

It almost sounds like hte issue we get with shift registers where LV get confused at to what type of data it should contain.

I can offer nothing else.

Ben

[X.]

Yeah, well, I was hoping that it would catch the attention of one of the blue guys...

It might not be a DVR bug, just one of these oddities of the Error List algorithm, as in this case...

Related to DVR reference not really DVR per se, BTW.

[Jon_S. · Accepted solution]

The error message for the DVR is a bug.  It will go away when the other errors in the VI are fixed.  CAR 295415 in the Known Issues.

[RavensFan]

Jon,

Something is wrong with the Added dates in that link.

Instead of a date, that CAR shows "Added: 00/40774/"

[Jon_S.]

I saw that when I linked the page.  I let the right people know internally to get it updated.  Our database automatically outputs Excel files that we can use to generate the page and it looks like some formatting changed.

The CAR above was added August 19th 2011 in case anyone was interested.

[mikeporter]

Actually, I have seen the same issue many times "breaking" a simple control reference when there was any sort of error in the VI. This is very intermittant.

Mike...

(Another one of the last ten people on the planet that knows what an octothorpe is...)