I am trying to run an already existing chunk of R code in a sandbox-ed fashion, by saving the global environment (containing functions and data) to a file, then loading it into a new environment (not the global environment) and evaluating a function call within that environment. However, I'm running into trouble with functions calling other functions in the environment. Here's an example:
f1 <- function(x) x*2
f2 <- function(y) f1(y) + 1
save(list=ls(), file="env.RData")
rm(list=ls())
jobenv <- new.env(parent=globalenv())
load("env.RData", envir=jobenv)
expr <- quote(f2(3))
eval(expr, envir=jobenv)
which fails:
Error in f2(3) : could not find function "f1"
whereas attaching the environment first works:
> attach(jobenv)
> eval(expr)
[1] 7
Is there a way to successfully evaluate a nested function call in the deserialized environment other than using attach?
(This question was forked from R - Evaluate a nested function in an environment based on commenter's suggestion)
