expect_equal_to_reference default behavior changed

[brodieG]

With the transition to expect_known, I think there was a regression introduced with respect to expect_equal_to_reference :

 #' @export
-#' @rdname expect_known_value
+#' @rdname expect_known_output
 #' @usage NULL
-expect_equal_to_reference <- expect_known_value
+expect_equal_to_reference <- function(..., update = FALSE) {
+  expect_known_value(..., update = TRUE)
+}

Should be:

expect_equal_to_reference <- function(..., update = FALSE) {
  expect_known_value(..., update = update)
}

Happy to submit a PR if you want, but it may be easier for you guys to fix. Let me know. For reference, the NEWS entry suggests the change is not intentional:

expect_known_value() replaces expect_equal_to_reference(), which has been soft-deprecated. It gains an update argument defaulting to TRUE. This changes behaviour from the previous version, and soft-deprecated expect_equal_to_reference() gets update = FALSE.

[hadley]

Doh. A PR would be great, although it's already been submitted to CRAN 😢

[brodieG]

Oh well, will submit PR later this week then. Still have to get my update to unitizer out so that it works with testthat 2.0.

[jimhester]

closed by #687

[jrcpulliam]

This bug seems to have reappeared somewhere along the line. I'm using 2.0.1, just reinstalled from CRAN. testthat::expect_equal_to_reference yields:

function (..., update = FALSE) 
{
    expect_known_value(..., update = TRUE)
}
<bytecode: 0x10d0ebd70>
<environment: namespace:testthat>

I'm switching to expect_known_value() but thought I should flag this.