10-11-2021 01:16 PM
Unfortunately I won't be able to upload files, but hopefully I can describe the issue in enough detail to get pointed in the right direction.
Teststand 2017, LabVIEW 2017 SP1 (17.0.1), both x32 versions on Win10 (x32 and x64 OS show same issue, multiple PCs)
NI Scope v20
All my drivers are behind a HAL using lvlibp in the runtime engine, only the scope calls to niScope Fetch Measurement.vi are affected thus far...
Only once have I seen an error message which mentioned a 'memory manager' issue (LV18 interestingly enough) - the rest of the times teststand just crashes after a few seconds when hitting the measurement step. Reproducible with an NI max simulated NI5114 instrument on my dev machine just like on the test station with actual hardware.
After trying various things to 'force' the call to use the 2017 RTE, stumbled on an odd 'fix' - renaming the LV RTE folder's to _2018 (and 19, 20)
Have tried manually & programmatically forcing the the 17.0.1 RTE when the test starts, as well as 'autodetect', and the lvlibp does not have the "allow future versions of LabVIEW to load this packed library" enabled, including other packed libraries with classes that my scope inherits from...to no avail. If the >2017 RTEs are available, it appears to try them and crashes.
With those runtime paths broken, the code works just fine from TestStand
Unfortunately the scope SFP in v20 requires the 2018 RTE, and renaming the folder requires admin privileges...so it's not an entirely palatable 'fix' for multiple test stations where testing would require the rename, yet calibration cycles (not to mention troubleshooting) would require a viable RTE.