LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Using LV Class Property Nodes in In-place Element structure causes the code to halt

I have some code written in LV RT 2010 that uses property nodes from a LVOOP class that I created.  I am using access the property node from inside an in-place element structure.  When the code reaches the property node, LabVIEW stops running.  (as if the abort button was pressed).  I verified this behaviour using highlight execution.  When I remove the in-place structure code runs fine.

 

Is this is know bug in LabVIEW?  Is there a CAR to resolve this created for a future release?

Message 1 of 8
(3,149 Views)

Hi Bob,

 

Can you attach an example or screenshot of how you are using the property node accessor in an in-place structure?  Are you accessing a property of an object that is contained in a different cluster or class?  An example will help clarify what causes this behavior.

 

Chris M

0 Kudos
Message 2 of 8
(3,137 Views)

I have remove the in-place function and still get the same problem.  Debugging is off in this VI.  The crash happens everytime there trigger list array increase for the first time.  If the array was zero, then become size=1 then it crashes.  If I run the code again it is fine (I assume since that memory is already allocated).  But if the array increases to size=2 then it crashes again.  Any help would be appreciated.  I am going to either remove the classes is initialize the array to some arbitrary size and the reduce it after the first run 

 

classes crashing.png

0 Kudos
Message 3 of 8
(3,118 Views)

Removing the property nodes for the class and replacing them with member access VIs seems to fix this problem.

0 Kudos
Message 4 of 8
(3,094 Views)

Bob,

 

I tried to reproduce the behavior you're seeing by recreating the Trigger and Trigger List classes, but I have not been able to reproduce it.  Please take a look at my version of the project in the attached zip file.  Take a look at Test.vi, and feel free to modify and repost it to show me how to make the problem occur.  I would like to file a CAR to make sure we get the problem fixed, but I will need to be able to reliably reproduce it first.

 

On a similar note, you can get much better performance out of the Process Triggers VI by preallocating the list array, instead of using build array in a loop.  Take a look at Process Triggers2.vi in my project.  You can run Benchmark.vi to see the performance difference between the two different approaches.

 

Chris M

Message 5 of 8
(3,077 Views)

Did you run this on Real-time?  I ran this function on my RT system and it did not crash I will try to send you the code I used to create the screen shot so you can see if that crashes?

 

Bob

0 Kudos
Message 6 of 8
(3,010 Views)

Bob,

 

I've run this on both Windows and RT (cRIO-9022) with no problems.  Are you seeing the problem only on RT, or does it happen for you on desktop as well?

 

Chris M

0 Kudos
Message 7 of 8
(2,987 Views)

I've seen this only on RT.  It is very hard to re-produce as it doens't happen every time for every VI but did happen in this one project.  We've since re-written this to remove the property node.  I think there is a Bug report for this issue now for 2010.

0 Kudos
Message 8 of 8
(2,763 Views)