Channel Wire Reference Leak

[irfanabid]

Hi,

I am elaborating a channel wire tag to stop parallel loop to make my code simpler as compared to notifier. However, I found a reference leak when using desktop execution trace toolkit. Any idea if I make any mistake in the program?

I am using this simple program to test.

Here is the message from Desktop Execution trace toolkit:

676413

21:40:14.1657490

Tag-bool.lvlib:))Channel.vi:7310015

Reference Leak

0

1

VI

Unknown Leak : 0x18067E78

Details: Unknown Leak : 0x18067E78
Call Chain:
-- Tag-bool.lvlib:))Channel.vi:7310015
-- Tag-bool.lvlib:Instantiate.vi:4380002
-- Untitled 3
VI Path: C:\Users\Engha\Documents\LabVIEW Data\2018(32-bit)\ExtraVILib\ChannelInstances\Tag-bool\))Channel.vi
Thread ID: 0 --- CPU ID: 1

[jkodosky · Accepted solution]

Your program is correct, and there is a reference leak-- we create a reference internally in order to enable probing the channel and we neglected to close it.  The fix will be in LabVIEW 2020.  Thanks for spotting this and alerting us to it.

--Jeff

[Ben]

---

@jkodosky wrote:

....  The fix will be in LabVIEW 2020. ...

--Jeff

---

There is only one person that has the intestinal fortitude to predict which version a bug will be fixed in.

Ben

[drjdpowell]

Question: is this an actual reference leak or just a nominal one?  In other words, could one construct code where an unlimited number of such references are generated?  Because not everything called a leak by the trace toolkit is actually a leak.

[Konan__]

---

@jkodosky  ha scritto:

Your program is correct, and there is a reference leak-- we create a reference internally in order to enable probing the channel and we neglected to close it.  The fix will be in LabVIEW 2020.  Thanks for spotting this and alerting us to it.

--Jeff

---

Welcome to Agile development!

[AristosQueue (NI)]

---

@drjdpowell wrote:

Question: is this an actual reference leak or just a nominal one?  In other words, could one construct code where an unlimited number of such references are generated?  Because not everything called a leak by the trace toolkit is actually a leak.

---

Merely a nominal leak. In fact, it isn't exactly a leak... we are leaking "not a refnum". What's being reported as a leak is the cleanup proc is running to cleanup what is *potentially* a refnum, but turns out not to be.

No, you cannot construct an infinite leak with this.

[drjdpowell]

I wish the trace toolkit labelled things potential reference leaks.