"could not determine data types" for TypeDef Error cluster to Error Case

[thols]

Just updated from LV6.1 to 8.0.1. I get the error "could not determine data types of all wires in this VI" for some of my VIs I am converting to LV8. The error is not highlighted, the front panel of the VI is just activated. However, I tracked the error to an error cluster that is type defined (for some reason) and connected to an Error Case. It seems LV6.1 accepts type-defined error clusters being connected to an Error-case, but not LV8. That is OK for me, I dont need the error-cluster to be type defined. But why does LV change from a case structure to an Error-case structure then, and why doesn't LV find the error and highlights it for me? I will solve this by disconnecting my error clusters from type-defines.

I haven't studied the upgrade notes thoroughly, but the behaviour by LV seems odd.

Reproduce like this:

Make an error cluster type-defined. Create a case-structure. Connect the type-define to the case structure.

Check the attached zip-file for VI and picture.

Download File:post-2740-1152103309.zip

[crelf]

This is a bug that (as far as I remember) will be fixed in the next version of LabVIEW - you need to manually recompile the VI (just hold down ctrl+shft and click the run arrow).

[thols]

Thanks, but recompiling won't work. I found some other post somewhere with the same error, but not the same cause, and recompiling worked for him I think.

[crelf]

It won't work, or it didn't work? Did you at least try it?

[thols]

Yes I have tried it, and I tried it again now. I mean't that it didn't work, sorry for the misuse of "won't". So, you mean it worked for you to recompile?

[crelf]

Ah - I understand now. I just downloaded your example and get the same problem (recompiling doesn't work). I get in contact with NI support if I were you - looks like a real bug...

[thols]

Contacted NI. I will post an update when I receive an answer.

[Grampa_of_Oliva_n_Eden]

Contacted NI. I will post an update when I receive an answer.

This sounds familiar.

If you change the constant to a control does the problem go away?

Ben

[Yair]

If you change the constant to a control...

That sounds like an LV 7.0 problem. Do you remember seeing it on other versions?

[Grampa_of_Oliva_n_Eden]

That sounds like an LV 7.0 problem. Do you remember seeing it on other versions?

Hi Yen,

When you get to my age you are happy with rembering things and being able to distinguish between memories and dreams.

Thanks for confirming that this is not just a dream. :thumbup:

I can not comment on what version of LV.

Ben

[crelf]

I can not comment on what version of LV.

I tried it on LabVIEW 8 and got the same problem...

[thols]

If you change the constant to a control does the problem go away?

No, the problem does not go away by changing to a control.

This problem does not exist in LV 5.1.1 and 6.1 (The code I found this problem in was originally made in 5.1.1 and later recompiled in 6.1).

[Yair]

No, the problem does not go away by changing to a control.

I was refering specifically to this (constant vs. control) as being an LV 7.0 only problem. It probably had something to do with a change to the way LV did constant folding between 6.1 and 7.0. As you can see, this seems unrelated.

[LAVA 1.0 Content]

I received the same error today (8.0.1).

My vis have been in a project (including an lvlib). An associate of mine attended Crystal Drumheller's "Demystifying the LabVIEW Project" session at NIWeek 06, and she suggested that a project should only be created after the VIs are 'mature'.

I'm still prototyping, and the rest of my group has not started using projects yet, so I decided to work on the VIs outside the project environment (didn't want to create 'issues' later)

I opened the project, removed all the VIs, controls, etc. from the project; leaving only the project name and the lvlib. I saved the project and exited. All is well... I then opened the higher level VIs and saw the "could not determine data types" error. I searched the LAVA forum and found the exact error AND the solution here (recompile).

I stopped, put on my LAVA button , and proceeded...

My problem was that ALL my typdef'd enum constants on the block diagram reverted to the first value . Also, all the typdef'd enum constants used as inputs to my sub-vis reverted to the default as well. It took me some time to check every one of these.

I THINK I dragged and dropped the typdefs from the project window to the block diagram, and also copied these on the diagram, but I'm not sure. I only post this in case someone else is considering "un-project-ifying" your code.