11-07-2006 11:53 AM
11-08-2006 05:22 AM
11-08-2006 06:39 AM
Hi,
Thanks for the response.
I was certain this was not the problem but added the deploy variables as you suggest. This did not improve the situation.
The problem is definately to do with the shared variable though. If I use a 'diagram disable' structure around the shared variable, everything works fine. Similarly, if I address the shared variable using datasocket, everything works. The problem only exists when I use explicit shared variables (the ones that look similar to local/global variables).
11-08-2006 09:33 AM
11-09-2006 03:15 AM
It sounds like this is a bug as I would have thought a LabVIEW support/system file would have been pulled in automatically.
I actually hit a bigger problem. I didn't want to include the test steps in the build script to allow me the flexibility of adding test steps without rebuilding the EXE. This does not work anyway as my TestSteps contain sub-VI's. It appears that when the TestStep is run dynamically, the connection/path to the sub-vi's is lost even if they are stored in an identical directory structure.
Thanks for your help.
11-09-2006 03:53 AM
11-09-2006 11:18 AM
James,
You are correct that you can call a sub-vi in a seperate directory and these paths are maintained as you say. The problem actually shows itself if you have one of your subvi's calling one of the NI library files. For example, add Clear errors to one of the sub-VI's that you were calling and you will see what I mean.
This does make sense I guess. Not having the dynamic VI in the build script means that any support files that they call are not automatically added. I could get around this I suppose by saving all of these support VI's within my project but I do not have the energy!!
Thanks
Colin