This VI Kills LabVIEW

[jon_mcbee]

When I run the attached VI (LV 8.6) it gives me the following message:

This message seems to  be thrown when the program enters the For Loop.  I have re-written this from scratch twice now and the same error is thrown at the same place.  I want to see if anyone else can reproduce this on a different machine, and if so if anyone has any ideas how to fix it.  I am posting it in 8.6 as this is the version I am having troubles in.   Any thoughts are welcome!

Message Edited by jmcbee on 10-29-2008 09:40 AM

[altenbach · Accepted solution]

I can reproduce this too. There is definitely something wrong.

Possible Workaround:

If you add real data to "array in" so it is no longer empty, it does not crash.

If you then make the "array in" empty again, it still works fine. 

Go figure. 😉

[Andrey_Dmitriev]

Hi,

That is nice... I was able to reproduce this, and simplified the code up to following construction:

The shift register is problem.

Andrey.

Message Edited by Andrey Dmitriev on 10-29-2008 05:01 PM

[jon_mcbee]

Kudos to both of you.  It seems that initializing a shift register with an emtpy array was not a problem in previous versions of LabVIEW.  I wonder if this is a bug or on purpose?

[altenbach]

I have added a note to the monthly bug thread so NI will investigate.

[altenbach]

A better code alternative would be to slice out the first colum and then autoindex at the loop boundary.

(This eliminates carrying the entire array slices inside the loop and further indexing at each iteration).

As a nice side effect, it no longer crashes, even with empty inputs. 🙂

Of course, the bug still needs to be fixed. LabVIEW can never crash, even with sub-optimal constructs 🐵

Message Edited by altenbach on 10-29-2008 09:22 AM

[jon_mcbee]

Goodeye.  Sometimes I get too caught up in the endgame and don't pay enough attention on how best to get there.

[jon_mcbee]

While I have an audience... The goal of this program is to take a 2D array (concentration in column 0, direction in column 1), and find peaks in the concentration column.  A peak must consist of atleast 5 points ( a peak with atleast two data points on either side, where the minimum point on either side is at most 50% of the peak).  Of course there can be multiple peaks in the array and since we are dealing with data taken on a circle there can be half a peak at 0-10 degrees and half a peak at 350-360 degrees, this is why I tack three copies of the array together.  If anyone has an elegent way to solve this problem I would love to hear it. 

Thanks!

Im going to start a new thread for this one, sorry about that.

Message Edited by jmcbee on 10-29-2008 11:13 AM

[Henrik_Volkers]

---

altenbach wrote:
[...]

Of course, the bug still needs to be fixed. LabVIEW can never crash, even with sub-optimal constructs 🐵

---

LabVIEW can never crash ??   

 

Message Edited by Henrik Volkers on 10-30-2008 09:00 AM

[altenbach]

In an ideal world ....

(Sorry, forgot the first few words. :D)