Why would LVCompare.exe show a non-cosmetic difference with attached vi labels?

[LTR_AZ]

LV23: I created left vi copied it to right so should be the exact same file except for the name. Opened each and ran VI compare.  Shows attached label differences even though it is a copied file.  Is there a reason why?

This adds a lot of chaff to comparing since every attached label is counted as a difference.

[LTR_AZ]

Working with support to test and they thought maybe the pixels moved...  So thought I would retry.

I created a single VI called "VI with attach label Orig.vi"
I copied the file at the file system level twice called:
"VI with attach label Orig_left.vi"
"VI with attach label Orig_right.vi"

I compared with Beyond compare in Binary mode.  All files are identical.

At Command Line:
C:\Users\nxf99929\Downloads>"C:\Program Files\National Instruments\Shared\LabVIEW Compare\LVCompare.exe" "VI with attached label Orig_renamed_left.vi" "VI with attached label Orig_renamed_right.vi"

Shows a difference in files that were initially the same at a binary level.
This matches what I am observing in my codebase.  Attached labels always show this difference.  It is not filtered by the cosmetic selection either.

[Kyle97330]

Can you install LabVIEW 2025 and check and see if the version of the compare tool that comes with it is any better?

If you look at its release notes, you can see there are noted changes to the compare tool.  A colleague of mine who tried 2025 said that he got less useless info on the compare tool when he tried it, but I haven't personally given it a go yet.

I personally suspect that the current compare tool is just using a quick "equals" comparison on all the properties of an object, and in this case it might be looking at the value of the pointer (0xABD153D0 or whatever) instead of what it should do, which is to check to see if the UID of the attached object matches.  And since it has loaded 2 copies of the VI into memory at the same time, their pointers are different.

[LTR_AZ]

Support was using LV2025 and claimed that the tool showed no differences in the diff tool.  So sounds like was fixed which is good just can’t migrate to 2025 quite yet. 

They are going to try LV2023 to reproduce on Monday.