`read_to_test` and `save_to_test`

[boennecd]

The read_to_test and save_to_test can be replaced by expect_known_output and expect_known_value as follows

# xyz.R
xyz <- function(verbose){
  if(verbose)
    cat("some output to the console\nin \nmultiple \nlines\n")
  1:5
}
# end xyz.R

# testthat/test_xyz.R
test_that("zyx gives correct output", {
  expect_known_value(xyz(FALSE), "xyz_results.RDS", update = TRUE)
})

test_that("zyx gives correct output", {
  expect_known_output(xyz(TRUE), "xyz_output.txt", update = TRUE)
})
# end testthat/test_xyz.R

# run 
source("xyz.R")
setwd("testthat")

# then run the the code in `test_xyz.R`. It creates the following files
readRDS("xyz_results.RDS")
# R [1] 1 2 3 4 5
readLines("xyz_output.txt")
# R [1] "some output to the console" "in "                        "multiple "                  "lines" 

# and the tests pass when you run 
setwd("..")
test_file("testthat/test_xyz.R")
#R √ | OK F W S | Context
#R \ |  2       | 0
#R == Results =====================================================================
#R OK:       2
#R Failed:   0
#R Warnings: 0
#R Skipped:  0

The update argument should be changed to FALSE after the update to ensure that test are not updated if they fail.