LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Which Wait-on-Notification's (of all types) are currently Waiting?

Solved!
Go to solution

I may have a wait-on-nofication deadlock. Before I use the Navy method (beat to fit, paint to match) by changing a bunch of WoN's Wait-on-Notification-with-History, it would be great simply to be able to see which WoN's are currently waiting. In this way I could maybe see which WoN's were participating in the deadlock and hence rethink my dataflow pattern. I haven't found a tool highlight WoN's, list WoN's, etc. But it sure would be handy.

 

Any info or ideas?

0 Kudos
Message 1 of 5
(2,715 Views)
Solution
Accepted by topic author evan1138

The desktop execution trace toolkit might show you who's waiting (or at least in which VI). I haven't used it so I'm not sure.

 

Another option for handling this in code is to use a timeout and not do anything if the timed out output is true, but this is most likely going to be a hack instead of a real solution in your case.


___________________
Try to take over the world!
Message 2 of 5
(2,710 Views)

It seems like you could also detect which WoNs are using the timeout, as tst said. Your "timed out?" output could just go to an indicator for debugging purposes. If your WoNs are in loops, you could set a small timeout and thus update that indicator more frequently.

Colden
Message 3 of 5
(2,689 Views)

@Colden R wrote:

It seems like you could also detect which WoNs are using the timeout, as tst said. Your "timed out?" output could just go to an indicator for debugging purposes. If your WoNs are in loops, you could set a small timeout and thus update that indicator more frequently.



Yes, thanks, I'm trying a variation on that now.

0 Kudos
Message 4 of 5
(2,681 Views)

I finally had time to deal with this bug, and using timeouts and errors with messages I was able to "ladder" up to find the deadlocked WoN's. Tedious, and I ultimately haven't solved it yet, just inserted a workaround for now. Thanks to all. The solution goes to tst who was first with the timeout suggestion.

0 Kudos
Message 5 of 5
(2,661 Views)