r_process subclass shouldn't be cloneable

[ElianHugh]

Since upgrading to R 4.4, I've been having a warning pop up every time I try to use callr::r_bg. I suspect it's a result of me updating R6 -- perhaps a change in R6 at some point changed cloning rules or warning behaviour?

callr::r_bg(\() {})
#> Superclass process has cloneable=FALSE, but subclass r_process has cloneable=TRUE. A subclass cannot be cloneable when its superclass is not cloneable, so cloning will be disabled for r_process.
#> PROCESS 'R', running, pid 187359.

^{Created on 2024-06-04 with reprex v2.1.0.9000}

Very small bug, but it can be annoying when logging or testing to be spammed with the R6 warning. I think the solution would be to change the r_process class to explicitly say cloneable=FALSE, though I'm surprised that cloning behaviour isn't implicitly inherited.

[gaborcsardi]

I don't think this happens with the released version of R6, but yeah, we should change it.