JuliaCall
JuliaCall copied to clipboard
Non-Contradiction
Question: Stop Julia session
No issue, just a simple question, maybe an idea for package extension:
Does it make sense to implement a function which terminates a Julia session? Are the ways to do this already ?
Thx
Currently there seems to be no way to stop a julia session. Are there any specific use case for this? Thx for the feedback!
This would be useful for my applications that start Julia processes that can run for days at a time. It would be nice to have a way to kill the process without having to close R.
Currently there seems to be no way to stop a julia session. Are there any specific use case for this? Thx for the feedback!
My usecase: I am iterating over many cycles via purrr everytime starting a new Julia instance. I fear this is not memory efficient.
@instantkaffee Unless you are starting a new R instance, otherwise JuliaCall will (nearly) always use the existing Julia instance instead of starting a new one.
@jjlynch2 There is still no general solution to stop Julia process in JuliaCall. One idea is to insert R callbacks into the julia function or statement you are executing. For example,
library(JuliaCall)
julia_setup()
jlongcalc <- julia_eval('longcalc() = begin i=0; while(true) i = i+1; RCall.rcall(:print, i); end; i; end;')
In the example, jlongcalc is a Julia function that executes forever by using while(true). However, in the function, an R callback is inserted RCall.rcall(:print, i);. So when the Julia code executes the R callback part, it will give the control back to R, and you can stop it just like stopping any R code.
For example, in RStudio, if you jlongcalc(), and click the stop button, then the execution of the function jlongcalc() will be stopped just like any R code.
PS: A similar idea is also used in C/RCpp when you have long calculations: if you want the C/Cpp code to respond to R stopping signal, then you need to insert some R callbacks into the C/RCpp code.
One disadvantage for this method is that it is very hard to become a general solution. If JuliaCall automatically inserts R callbacks into users' Julia code, it will make the users' Julia code running slower than it should be.