search directory configuration before creating deployment file

[AnandR]

Hi,

when we creating the deployment file using teststand deployment utility , i have some question on that.

1. Is teststand deployment utility include all the files enabled in search directories.

2. Do i need to specify the process model file in the search directories, will the teststand automatically include the process model.

3.How to reset the search directories to default value.

Regards,

Anand

[~jiggawax~]

Which version of TestStand are you using?  In 2014 and earlier the search directories are stored in TestExec.ini.  In 2016 and later they are stored in SearchDirectories.cfg.

The deployment utility does use the search directories.  Basically if you can't run your sequence file in TestStand then there's a chance it won't deploy correctly.

You can reset to default search directories by deleted either TestExec.ini (2014 and earlier) or SearchDirectories.cfg (2016 and later).  Then restart TestStand.  It will create a new one of these files.  I would caution that if it's TestExec.ini there are a whole bunch more settings in there than just search directories so be warned.  Thinks like station settings and such.

If you are using an out-of-the box process model then you would need to specify to include it.  The TestStand help discusses how: https://zone.ni.com/reference/en-XX/help/370052R-01/tsdeploysystem/infotopics/deployengine/

Hope this helps,

[AnandR]

Hi Jigg,

I am using teststand 2016

i am using sequential process model, teststand simple userinterface customized for my application and my application(client) sequence file and its code module.

When i start creating deployment files , i got a question like how to organize the client sequence file, code modules , user interface code and sequential process model in my  directory to feed into the deployment utility to create deployment file. 

I am bit cautious here since , earlier tried to use the deployment utility and it modified all my labview c: directory which made me to reinstall labview again.

Regards,

Anand 

[~jiggawax~]

It sounds like you might need to take a course on proper TestStand deployment.  The deployment utility should never touch your source code.  It makes copies of your source and then re-links them to the new locations of dependencies.  Also, you should never install a deployed TestStand project on your development machine.  I could see how that could easily destroy your source code.  Especially if you are deploying to the same place as your source code.

Have you read this: https://zone.ni.com/reference/en-XX/help/370052R-01/tsdeploysystem/infotopics/identifyingcomponents/

Kristen made a pretty good video here: https://www.youtube.com/watch?v=E04KPyWrEuo

Here's a non-NI read on it: https://virinco.zendesk.com/hc/en-us/articles/115001244571-TestStand-Deployment-guide

Hope this helps,

[AnandR]

Thanks Jigg.

Regards,

Anand

[AnandR]

Hi Jigg,

what are the search directories should selected before we start the build in teststand deployment.

I have a shared sequence file. it does not have main sequence and it has labview code modules.

The sequence in this file is used in sequentialprocessmodel.seq. 

Before deploying where should i keep the sharedsequence file. DO i need to add this in search directory.

Regards,

Anand

[~jiggawax~]

The search directories should be able to find all of your dependencies that are called directly from your sequence file.  The easiest, and recommended, way to do this is to make all of your paths in the VI calls to be relative to your sequence file.  The deployment utility will re-link them correctly depending on where you are deploying to.  This being said, I don't do it that way.  I like the control of the search directories and so I just add the top level folder to my VIs in my search directory and check the search subdirectories button.  Also because I share code with multiple developers and relative can be a bit tricky if for whatever reason our user name is in the path.

So ultimately there are 3 sets of the same code when deploying.  We will call them Source Code, Image Code and Deployed Code.  Your source code should reside in the location that you are developing out of.  For most developers this is a source code control (e.g. SVN, GIT HUB, ClearCase, etc...).  However, for novices it is usually a thumb drive labeled MyTestv2.5.zip (this is a jab at people that don't know how to develop code).  TestStand search directories should be pointed at your source code, wherever it may be. 

The deployment utility then takes the source code and creates an Image.  The Image code is basically your source code but moved to one file hierarchy and relinked to the relative paths within that file hierarchy.  The Image is used by the Installer (if you are creating one).

Then you either run the installer or copy the Image to the deployment computer (note this should not be done on the development computer unless you know for certain what you are doing).  Once the code is on the deployment computer it becomes the deployed code and is in a state where it can be executed.

So if we look at this simple breakdown of how deployment works and apply it to your shared sequence file then basically you develop it on your development machine in a source code location.  Because it is used by the process model I'm assuming it's in the TestStand public folder.  Either way it doesn't matter as long as your search directories can reach it.  Then you add this shared sequence file to your tsd in the deployment utility (either through your workspace or a directory).  Once you build from the deployment utility your shared sequence is put into the image and relinked with its dependencies.  Then you move it to the deployed machine either via installer or via copy image to new machine.  On the deployed machine the search directories need to point to the new location.

Hope this helps,

[AnandR]

Hi Jigg,

Thanks for your well written answer.

I am trying create the teststand framework deployment for the first time. Just thought to ask this question again.

1. Currently my application sequence file are in 😧 drive(whichever, ) . But not in C: drive.  Do

 i need this to move C: drive inside C:\Program Files (x86)\National Instruments\TestStand 2016.

2. I took the simple user interface provides by teststand and i customized it as per my requirement.

Currently it is in my 😧 drive.  DO i need to move this C:\ drive before deployment.

3.I am using sequential process model and also i customized as per my requirement. This process model file uses the labview modules and currently these files are in my d: drive. Do i need to move to C: drive.

Overall i am not still clear on how to organize application sequence files , labview modules , process model file and its labview modules , User interface code. in order to create deployment file.

Regards,

Anand

[~jiggawax~]

So if you think about the 3 states of your code during deployment: Source, Image, Deployed.

Source: does not matter where it resides as long as the search directories can reach it.

Image: does not matter where you put this.  It should be temporary and search directories should never see this.

Deployed: Very Important where this goes on the deployed pc.  Should not deploy back to your development pc, especially if deploying to the same locations as your source code.  This can cause major linking issues and potential loss of source code.  The search directories on the deployed pc should be able to point to the deployed code.

Hope this helps,