LabVIEW confuses subVIs in different paths but have same name?

Question

I am using some of the subVIs that other people wrote so this is some historical issue.

The problem I have is that if two different subVIs are different, stored in different locations, but happens to have the same name.

It seems sometimes LabVIEW confuses the two of them and the link jump from one to the other, since the two subVIs are different, this makes errors.

So my question is: is this something real? That LabVIEW call subVIs by their filename rather than absolute filepath. Or am I describing something unreal and may be caused by other issues.

Thanks!

Answer 1

There are two separate things here. The first is that LV remembers where each file that something uses is placed, and there are three ways it can store that path (assuming I'm not forgetting any):

1. Logical. These are folders like vi.lib, where the base folder is represent by a name (so the path is stored as "|vilib|\...\myVI.vi"). This allows you to call the correct file when you go to a different LV version.
2. Relative. This is how LV normally saves paths, so if VI A calls VI B and they're in the same folder, the path to B inside A will be stored as "..\B.vi".
3. Absolute. I believe this is only reserved for VIs which are stored in a different volume, but do not match the first rule.

If for whatever reason LV doesn't find the file where it's supposed to be, it will take the one already in memory or try to search for it in other places, at which point you get a searching dialog.

The second thing is that there can only be one VI with a specific name in memory at any one time. If you load two VIs with the same name, you will get a conflict and have to resolve it, as you've seen. This isn't LV getting confused, just a matter of each calling VI referencing another VI with the same name and the order in which you load them. Technically, I suppose that LV could have used the full paths to allow duplicate names, but I believe that this decision goes back at least 20 years, so it won't be changed now, and it certainly would be a problem when you consider that now you can reference VIs in memory by name. It also has some other technical merits.

If you do have VIs which need to have the same name, what you will most likely want to do is namespace them. In the past, this was often done by actually changing the name using a prefix or a suffix. These days, you would usually do this by placing the VI in a library (such as .lvlib or .lvclass). This doesn't modify the file name, but it does become part of its fully qualified name, which is what determines collision, so it allows you to have multiple VIs with the same name. In the case of classes, this is actually a requirement, as that's the mechanism used to determine which VIs will override.