LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Sever and dangerouse Issue usign Feedback node (68872 4EBC79AG or similar) still exist in LabVIEW 8.5.1

I've created the attached vi in LV 8.5.1. During editing an internal error popups and I became unable to save the vi. So I've copied and pasted the whole diagram to another vi and saved it. After a reboot I've reopened the saved vi and I've discovered that all the feedback nodes was messed up (init termial draw far away from the while loop). After few editing LabVIEW... suddenly disappeared!!!!!! No crash signaling, and then no recover of opened vi was possible!
 
This is a really dangerous BUG because the LV crashing can cause severe data/work loss.
 
My suggestion is to not use feedback node until a real fix wil be released.

Happy coding to all!
Regards
 
Message 1 of 15
(4,219 Views)
Ooops... others missing attachments: crash report.
Download All
0 Kudos
Message 2 of 15
(4,215 Views)

I don't see the problem you described.

I can save the file and open it up and I don't see any errors or crashing.

R



Message Edited by JoeLabView on 07-22-2008 07:35 AM
Message 3 of 15
(4,195 Views)

Of course you can open and save the vi but take a look at the position of the Initialization terminal of the feedback node (also in your snapshot one can see that something is wrong on the left side of the while loop)! Try to move them or try to move the various initialization constants.... strange things will happen, the worst is a LV crash!

Smiley Surprised

0 Kudos
Message 4 of 15
(4,174 Views)

Good observation..

lemme try..

0 Kudos
Message 5 of 15
(4,168 Views)
How did you manage to move your initializer terminals outside the While Loop?
 


Message Edited by JoeLabView on 07-22-2008 11:21 AM
Message 6 of 15
(4,165 Views)
When I use the feedback node, I can move the initialization node to the edge of the While Loop as you would expect.
I did see something that might be a bug in the graphics refresh when moving the feedback node up or down..  It doesn't happen all the time, but between 25 to 50 % of the time on my machine.
 
I could move your nodes back to the While Loop and the snapped to the appropriate locations (the other two as well).  It does give a broken arrow that indicates that the feedback node contains unwired or bad terminal.
 
I will try to "fix it" and post back your code.
 
R
 


Message Edited by JoeLabView on 07-22-2008 11:30 AM
Message 7 of 15
(4,160 Views)
Here is your VI.  Hope it is fixed.
 
Let me know.
 
R
 
 
Message 8 of 15
(4,156 Views)

Tried it again in another new blank VI..



Message Edited by JoeLabView on 07-22-2008 11:37 AM
Message 9 of 15
(4,150 Views)

Thank you for your attention!

I've opened the fixed vi and all seems to be ok, but in few mouse click all messed up again causing strange floating wire and a LV crash.

If you are brave please try the following procedure (save and close all of your work..!):

  1. Open your fixed vi.
  2. CTRL-E to display the block diagram.
  3. Rigth click on the upper (DBL) feedback node and change it to "Globally Initialize".
  4. Right click again on the same feedback node and select "Move Initializer One Loop Out".
  5. Select the feedback node, the "+", the "+1" and the other (blue, INT32) feedback node.
  6. Then perform the "Horizontal Centers" alignement.
  7. (in my case) voila! Move one of the 2 feedback node and see what happen....
  8. Now try to close and save the vi.

I run LV on a NI PXIe-8106 so I think that this kind of hw is fully compatible with NI software... 🙂

I suspect that in this case the problem is not at a graphic level but reflect a corruption of internal data LV data.

 

Download All
0 Kudos
Message 10 of 15
(4,129 Views)