Open llavall opened 3 years ago
No. This just creates extra work. What we need is to control properly every error (checked exception) in the runner so that the stack traces are not hidden and a meaningful error is shown in the console. This is something Thomas is already working on.
I still think the suggestion is a valid one. Unfortunately we do not have time for it now, so we must give it a lower prio
I have realised that each time an FSKX is failing to run it is giving the person who has to find out why a hard time. One way to solve that is to copy all important model parts out of the FSKX-file (which is not easy at all) in order to run that in the native environment of the model (e.g. Python or R).
Since we already have an FSK to R node, I was wondering wether it is an easy to accomplish option to allow the user easily to simply attach an R node for example and find the ready to run model in there and just run it.
This would allow easy bugfixing, could serve as a test if the FSK result is the same and may open other options for the user.
Best, Lars