Nasty LV Quirk

[paul_cardinale]

Here's a really nasty LabVIEW Quirk:

1. Create a custom control whose type is set to "TypeDef."
2. Place it on a VI's front panel.
3. Create a property node for that control
4. Save and close the VI
5. Change the custom control's type from "TypeDef" to "Control"
6. Reopen the VI.  The control will still be there, but the property node will be gone.

[wiebe@CARYA]

.Reproduced in LV20, 64 bit...

---

@paul_cardinale wrote:

4. Save and close the VI

---

This asked to save the type def as well, as I didn't save it before (as this wasn't instructed 😁). It still reproduced though.

[billko]

LV 2016 also.  I believe that, since the property node is an implicit one, LabVIEW thinks you just deleted the control so the implicit property disappears along with it.  I guess the typedef doesn't get "apply changes" to the control, so the implicit property node is orphaned and gets deleted.

[billko]

You can even see it in real time.  Leave the VI open and change the typedef to a control (and apply changes).  The property node disappears before your very eyes - POOF! - it's gone.

[altenbach]

Interesting edge case. I have never felt the need to convert a typedef back to a control.

Typically I would just "right-click...disconnect from typedef" and the property node is retained.

[paul_cardinale]

---

@altenbach wrote:

Interesting edge case. I have never felt the need to convert a typedef back to a control.

Typically I would just "right-click...disconnect from typedef" and the property node is retained.

---

We have a lot of custom buttons that always should have been controls instead of typedefs.  The link is unnecessary and just creates extra stuff on our deployments.  I switched them from controls to typedefs then lots of stuff broke.  I'm trying to think of a way to fix things without going back to typedefs; so far no luck.

[billko]

How about finding each instance and disconnecting it?  Then, only after all the typedefs have been disconnected, make the typdef a control?

[paul_cardinale]

I found that this will fix a VI, if it hasn't been opened:

But this code makes some assumptions that may not be true.  I'll have to refine it before I shotgun our code.

[paul_cardinale]

On second thought, I'll probably do something like this:

1. Make sure the test engineers have their latest code checked in.
2. Revert the buttons back to the typedef'd versions
3. Write and run a script that:
   1. Scans all our code for instances of those buttons
   2. Checks out the owning VIs
   3. Disconnects the typdefs
   4. Checks in the the VIs
4. Unrevert the buttons to the non-typedef's version.
5. Tell the test engineers to sync.

[billko]

---

@paul_cardinale wrote:

On second thought, I'll probably do something like this:

1. Make sure the test engineers have their latest code checked in.
2. Revert the buttons back to the typedef'd versions
3. Write and run a script that:
   1. Scans all our code for instances of those buttons
   2. Checks out the owning VIs
   3. Disconnects the typdefs
   4. Checks in the the VIs
4. Unrevert the buttons to the non-typedef's version.
5. Tell the test engineers to sync.

---

Oh, gosh - so kind of what I suggested on a grand scale... whew!