I have a ROOT built from source with the following options:
gnuinstall=ON
imt=ON
minuit2=ON
rpath=ON
pyroot=ON
I had problems using PyROOT because I didn’t source thisroot script but now I have a line in my configuration files and I am able to import ROOT module inside a Jupyter notebook.
The problem is I noticed that at every start of Jupyter Lab a ROOTSYS folder is created in the current path. I am wondering what is its purpose and if its creation can be avoided in some way.
for me it looks like a your are setting JUPYTER_CONFIG_DIR, that file “migrated” for examples is usually generated in the path ~/.jupyter but if you set it to your own path all that files will be generated there.
I try just sourcing thisroot.sh and I am able to use ROOT and not code is generated.
I found the problem: I use fish shell instead of bash so I source thisroot.fish instead of thisroot.sh; looking at the code I noticed the missing $ corrected by commit 0be8bbe on master branch but still not on latest-stable.
Adding the $ before ROOTSYS in thisroot.fish should resolve the problem.
Thanks for the help.
EDIT: this solution is not resolutive because I use a fixed location installation (gnuinstall=ON) instead of the more common location independent. So I have to correct $ROOTSYS/etc/notebook -> $ROOTSYS/etc/root/notebook too
Uhm with gnuinstall=ON, the assumption is that components will end up in standard system directories and $ROOTSYS won’t be needed at all. Feel free to open a bug report about the bad interaction of gnuinstall=ON and jupyter.