I have some amount of code in R left over from my past research. It's not very fast, but it works. Julia generously allows to postpone the full migration process for later, using this R code via Rcall, this is very convenient. Recently I decided to prepare one interactive demo using Pluto and among other things in one place used the insertion of R code via Rcall. At the first launch, if only open notebook, everything works, but if change anything in this code block and recalculate it, Pluto returns an error like this:
cannot assign a value to variable workspace407.b from module workspace408
This error occurs even when using the simplest code, so the problem is not an error in the R code, but in the relationship between the environments when updating cell with this code. If run this code directly from Julia, without Pluto, there is no such problem. Here is a completely primitive example that demonstrates the problem:
begin
a = 2
@rput a
R"""
b = a*2
"""
@rget b
b
end
For this to work, using Rcall must be called somewhere before and R must be installed and configured for external startup. On the first run of Pluto notebook with this code, everything works as it should. But if change for example b = a*2 to b = a*3 and update cell with this code, Pluto will give an error. How to solve this problem?