Package Builder Bitness

[RN888]

PC Details

Processor        11th Gen Intel(R) Core(TM) i7-11850H @ 2.50GHz   2.50 GHz

Installed RAM        32.0 GB (31.7 GB usable)

System type        64-bit operating system, x64-based processor

Edition        Windows 10 Enterprise

Version        21H2

Installed on        ‎20/‎08/‎2022

OS build        19044.2486

Experience        Windows Feature Experience Pack 120.2212.4190.0

  System summary

   Date               02/08/2023 12:58:55

   NIPB version       22.5.1.49816

   NIPB bitness       64-bit

   NIPM version       23.0.0.49275

   Operating System   Windows 10 64-bit

I have a test system coded in 32bit Teststand and LabVIEW.

Recently I have been forced to upgrade my PC, and have downloaded all my code from subversion. Including the package builder file.

I had a contractor create the package build file for me, and it has run fine for several months in the past.

However now I have the following questions

1

Why does the builder generate the error below multiple times. And yet succeeds in the build?
[12:58:55.28 pm] Warning 16049: The Package Builder support for TestStand failed to load and has been disabled in this instance of Package Builder.

  Unable to load TestStand engine. Make sure a supported version of TestStand is installed and the bitness matches that of Package Builder. […]

[1:01:46.18 pm] ======= Build Packages: 6 Succeeded, 0 Failed, 0 Canceled (elapsed time 00:04.32) =======
 

2

Is this something I can ignore as there are no errors in the report?

3

Due to the warning 'Make sure a supported version of TestStand is installed and the bitness matches that of Package Builder', I tried the 32bit builder

Why does the 32 bit package builder generate more errors? I would have thought it would be more compatible.
I have loaded the pbs file into 32bit Package builder. But it crashes. And generates a lot of 'references the missing package' type warnings as soon as the pbs file is opened

Ie  about 200 instances like  The package 'ni-license-manager' references the missing package 'cba7edb3-a142-43b6-b68b-a723624f1375' with a relation of 'Conflicts'

Then a lot of errors like

[1:48:35.04 pm] Warning 16511: The dependencies of the following TestStand sequence file step cannot be discovered because the step is specified by expression.

   File not found in step:   Call Model Plugin - Initialize

   Sequence file:            C:\+Files\CPE SVN ABB UI\Software\Multi DUT Test Platform\Workspace\TestStand Model\ModelSupport.seq

   Sequence:                 Initialize Model Plugins

   Step Name:                Call Model Plugin - Initialize

   Step ID:                  ID#:4vv6wUUUSEijXxOqZyPoDB

   Expression evaluated:     Locals.ModelPlugin.Base.RuntimeVariables.AbsolutePath

The crash occurs after about 20 of these warnings

Any help much appreciated. Thanks in advance.

[Scott_Richardson]

Hi RN888, the "Warning 16049" is typically added once to the Output log on launch of NIPB to clearly communicate that NIPB cannot handle TestStand files because the TestStand engine is not installed or cannot be loaded. This is so that users are warned when trying to build a package that contains TestStand sequence files that cannot be properly preprocessed by TestStand before being placed into the package.

Since the TestStand engine is an in-process COM server, the bitness of the NI Package Builder process must match that of the installed TestStand engine, so that is why there is a 64-bit version of the NI Package Builder to use with TestStand 64-bit and a 32-bit version to use with TestStand 32-bit. You should use the NI Package Builder bitness associated with the bitness of TestStand that you will be deploying. What version and bitness of TestStand are you wanting to use?

The "Warning 16511" occurs because the package included what appears to be a custom ModelSupport.seq file that dynamically calls a sequence file on disk using an expression, and NIPB cannot determine the file that it calls for inclusion in the package to ensure success on a deployed system.

Regarding "the crash occurs after about 20 of these warnings", I will have to try to reproduce here. In the mean time, can you please share the NIPB log file created, %localappdata%\National Instruments\NI Package Builder\Logs\*.txt that is associated with the crash? You can share to this thread or send to me in a private message.

[Edit] I did a quick test with NIPB 22.5.1 and TestStand 2021 SP1 and was able to add to a package all of the default model files and build the package successfully. I did get a lot of "Warning 16511" messages due to the nature of their design of calling sequences dynamically.

[RN888]

Thanks for the response Scott. Apologies for the delay. For some reason the post did not go through yesterday.

SR> You should use the NI Package Builder bitness associated with the bitness of TestStand that you will be deploying.

RN> I tried 32B Package builder with 32B Teststand, but the builder was unstable.

        Note that for some reason in the past the build package was run as 64bit and worked.

        I only have the header as evidence not the log file.

SR> What version and bitness of TestStand are you wanting to use? 

RN> 32bit, log attached.

SR> NIPB cannot determine the file that it calls for inclusion in the package to ensure success on a deployed system

RN> As you mentioned. I think the ModelSupport.seq is a custom one. I don't have much to do with it. It was written

        by contractors. I struggle with the  level of coding in there.

         The warning mentions.'Expression evaluated: Locals.ModelPlugin.Base.RuntimeVariables.AbsolutePath'
         This appears to be the common thread. I can't find in the code where it is written to. 

         I'm wondering if the absolute path is an issue because the development and deployment paths are different.

         I would have thought the paths would be resolved for the different environments. Though I can't see where.

[RN888]

Log file. Using my old working pbs file, and 32bit package builder

[Scott_Richardson]

RN888, as you indicated in the log file there was no indication of why the crash occurred. The log file that you sent appears to be the file next to the solution. Just wanted to double check that you sent me the corresponding log file in the %localappdata%\National Instruments\NI Package Builder\Logs\ directory because there is a slight chance there could be more information in that.

You could also check the Microsoft Event Viewer (Enter eventvwr.exe on the command line), and look at the "Event Viewer (Local) >> Windows Logs >> Application" item in the tree and find any corresponding Error events near the time of the crash, and share the text for those events which might have a stack trace.

The only other option might be if you want to take this offline and share the project and the model files with me to try to reproduce the issue here at NI.

[RN888]

Hi Scott.

Yes I would welcome the help.  I'll contact you by email.

Thanks Richard