>1) That the data structure changes actually happened. He sent me the two relevant tables after running the update code the other day and the new memo field is there; the old char fields are gone. Since he could find these where expected, I think I can also rule out something related to virtualization.

Looks like I was wrong and maybe this is a virtualization problem. I got remote access to their system and am testing bit by bit.

I've confirmed that the app is looking at the folder I expected for data, and that the table is the one I think it is. However, when I test for the memo field's existence, it's not there. Specifically, I added this code to the main program:

WAIT WINDOW "Data folder is " + m.g_data
USE ProdList
WAIT WINDOW DBF("ProdList")
WAIT WINDOW "ProdList " + IIF(TYPE("ProdList.mNotes") = "M", "has ", "does not have ") + "memo field."
USE 

g_data is a variable that holds the path to the data (retrieved from another table). It's the folder I'm looking in, and DBF("ProdList") is the file I'm looking at. I can see that there's an FPT (which wasn't there before I ran the update code), but the test above says it doesn't have the memo field. I copied ProdList.* to my machine and I can see the memo field is there.

So despite all evidence to the contrary, it appears to me that the application is not using the file I think it is. Here's the set-up.

Client machine is Win10 (but the app gave the same bug on a Win7 client earlier).
Server is Win7
The data is on a mapped drive and is not in the Program Files hierarchy.
To test, they've turned off virus scanner on both machines.

Is there a way to actually check whether a file is being virtualized? Also, why wouldn't the updated table be properly virtualized regardless?

Tamar