Building model for Linux RT target from VI

[beavercreek]

@Darin G,

Are there plans to clean up the process to install what is required to build Linux models?   Seems strange to install Compact RIO device drivers when I've never seen or used one and only use a PXI chassis.

Also could skip having to install and muck around with Eclipse when maybe creating/shipping a .reg file would suffice?

Also an update, re-installing real time on the first problem PC did not get it working on that one yet.  Just happy I have at least one machine capable of lvmodelso creation.  I'll have to diff their software packages and will be sure to update the thread if I learn anything.

[Darin_G]

Glad you're up and running... or at least moving forward.

We definitely see a fair amount of support issues related to model compiler toolchain setup.

Making this simpler is not our direct & immediate focus, but we have a few projects that will improve this throughout the VeriStand 2021 releases.

We see step one of fixing this as: simplify the matrix of targets we support so that we can better tailor the user experience. So starting with VeriStand 2021, we are dropping Phar Lap and Linux ARM RT support from VeriStand, which gives us 50% fewer model execution OSes to support. This will give us a more efficient route to simplifying the install experience for Windows & Linux RT going forward.

[beavercreek]

I checked that all software on the PC capable of building vi's into lvmodelso files is also installed on the PC that cannot build them.  I did notice that the PC that cannot do the builds has several versions of Real-Time installed (because of the several versions of LabVIEW that are also installed).

I re-ran back through the eclipse instructions to the point of successfully building the hello world project again, but still no joy on that PC for converting a very simple VI.

Do you think multiple versions of Real-Time could be responsible?  Or if that was the problem, then building for PharLap would fail as well?