The change in ggplot2 to use tidy evaluation means quosures are used inside of ggplot2. Thus, this class of unit tests...
g <- mcmc_intervals(arr, rhat = r)
rhat_map <- g$layers[[3]][["mapping"]]
expect_identical(rhat_map$colour, as.name("rhat_rating"))
#> Error: rhat_map$colour not identical to as.name("rhat_rating").
#> target, current differ in having response: FALSE, FALSE
will fail because the former is a quosure and the latter is a name.
rhat_map$colour
#> <quosure>
#> expr: ^rhat_rating
#> env: 000000001DE45EF8
as.name("rhat_rating")
#> rhat_rating
But as far as I can tell, we can get the name with
rlang::get_expr(rhat_map$colour)
expect_identical(rlang::get_expr(rhat_map$colour), as.name("rhat_rating"))
Here are the first batch of errors...
test-mcmc-intervals.R:89: failure: mcmc_intervals/areas with rhat
rhat_map$colour not identical to as.name("rhat_rating").
target, current differ in having response: FALSE, FALSE
test-mcmc-intervals.R:98: failure: mcmc_intervals/areas with rhat
rhat_map2$fill not identical to as.name("rhat_rating").
target, current differ in having response: FALSE, FALSE
test-mcmc-intervals.R:99: failure: mcmc_intervals/areas with rhat
rhat_map2$colour not identical to as.name("rhat_rating").
target, current differ in having response: FALSE, FALSE
test-mcmc-intervals.R:105: failure: mcmc_intervals/areas with rhat
rhat_map4$colour not identical to as.name("rhat_rating").
target, current differ in having response: FALSE, FALSE
test-mcmc-intervals.R:109: failure: mcmc_intervals/areas with rhat
rhat_map5$colour not identical to as.name("rhat_rating").
target, current differ in having response: FALSE, FALSE
== Terminating early =================================================
Too many failures
The change in ggplot2 to use tidy evaluation means quosures are used inside of ggplot2. Thus, this class of unit tests...
will fail because the former is a quosure and the latter is a name.
But as far as I can tell, we can get the name with
Here are the first batch of errors...