Creating a shared library with juliac

I would like to experiment with using R’s .Call interface to call Julia functions compiled into a shared library. See Using rcopy and sexp methods to create a .Call interface to a compiled Julia library · Issue #556 · JuliaInterop/RCall.jl · GitHub for more details.

I have not yet been able to compile a shared library and dyn.load it in R. My main development platform is an M1 MacBook and I am encountering the issue described in Warning on Mac M1 · Issue #2 · TheCedarPrince/InteroperableJuliaBinaries · GitHub I have also tried on a computer running Ubuntu and got an unresolved symbol. (I can provide more details if needed.)

Can someone point me to an example of successful creation of a shared library using recent juliac which I can try to imitate? The usual caveats apply - I know it is an experimental capability and under active development. However, I do think that easy ways to call Julia code from R will be a big boost for Julia adoption in data science and I am willing to work on the bleeding edge

2 Likes

See repo links in