How to use customize functions in the `func` function? about callr HOT 4 CLOSED

I am not sure what you mean by "quote". You mean you want to call other functions from the same package? Then yes, use the package argument.

from callr.

I am not sure what you mean by "quote". You mean you want to call other functions from the same package? Then yes, use the package argument.

yes, just use functions in the Global Environment, but callr says "could not find function"

> add_1 <- \(x) return(x+1)
> loop_x <- function(){
+   i <- 1
+   while(TRUE) {
+     i = add_1(i)
+     readr::write_lines(i, "./data/x.txt")
+     Sys.sleep(1)
+   }
+ }
> rtask <- r(loop_x)
Error: 
! in callr subprocess.
Caused by error in `add_1(i)`:
! could not find function "add_1"
Type .Last.error to see the more details.

in this case, I think the add_1 function is stored in the .GlobalEnv, where package argument by default directed to?

from callr.

Yes, that is by design, only the selected objects are copied to the subprocess. Copying the global environment is often a bad idea, if you have large objects, e.g. data sets in the global environment.

If you want to copy a set of functions, here is a way:

fun1 <- function() 1
fun2 <- function() 2
funcs <- list(fun1 = fun1, fun2 = fun2)
fun <- function(x, ..., funcs) {
  x + funcs$fun1() + funcs$fun2()
}
callr::r(fun, list(x = 3, funcs = funcs))

#> [1] 6

If you want to copy the whole global environment, you can do that as well, but again, it might be a bad idea:

callr::r(fun, list(x = 4, funcs = as.list(.GlobalEnv)))

#> [1] 7

from callr.

Thanks for your patient explain! Now I understand why the global environment functions can't be read.
And your solution is quite inspiring, I almost forgot that functions can be put into a list() 😀

from callr.