rlang warnings in ggplot2 unit tests

[clauswilke]

When I run the ggplot2 unit tests, I get two rlang warnings. Does anybody else see this? Is this a misconfiguration on my end or is something not quite right with the test or the code base?

==> devtools::test()

Loading ggplot2
Testing ggplot2
✔ |  OK F W S | Context
✔ |   1   2   | Adding plot elements
───────────────────────────────────────
test-add.R:4: warning: mapping class is preserved when adding uneval objects
Subsetting quosures with `[[` is deprecated as of rlang 0.4.0
Please use `quo_get_expr()` instead.
This warning is displayed once per session.

test-add.R:4: warning: mapping class is preserved when adding uneval objects
Subsetting quosures with `[` is deprecated as of rlang 0.4.0
Please use `quo_get_expr()` instead.
This warning is displayed once per session.
───────────────────────────────────────
✔ |  14       | test-aes-calculated.r
✔ |   9       | Aesthetics (grouping) [0.4 s]
✔ |  10       | Aes - setting values [0.5 s]
...

Strangely, I cannot reproduce these warnings in a reprex:

library(ggplot2)
library(testthat)

# the test from test-add.R, line 4
p <- ggplot(mtcars) + aes(wt, mpg)
expect_identical(class(p$mapping), "uneval")

^{Created on 2019-10-05 by the reprex package (v0.3.0)}

[karawoo]

I see this too when I run the tests locally, and also can't reproduce outside of running the tests.

[yutannihilation]

This seems because rlang:: signal_soft_deprecated() behaves differently when it runs in testthat environment.

#' * `signal_soft_deprecated()` warns only if called from the global
#'   environment (so the user can change their script) or from the
#'   package currently being tested (so the package developer can fix
#'   the package).

https://github.com/r-lib/rlang/blob/1f65d7c79315eee74b17991d289edeea6626467f/R/compat-lifecycle.R#L90-L98

[thomasp85]

Have anyone successfully identified where the calls occur? I can only provoke these during testing, not by replaying the tests in the console, and searching for [[ is pretty hopeless

[yutannihilation]

It says strip_dots(). So, probably here?

library(ggplot2)
library(testthat)

trace("[[.quosure", quote(lobstr::cst()), where = asNamespace("rlang"))
#> Tracing function "[[.quosure" in package "namespace:rlang"
#> [1] "[[.quosure"

p <- ggplot(mtcars) + aes(wt, mpg)
#> Tracing `[[.quosure`(expr, 1) on entry 
#>      x
#>   1. \-ggplot2:::`+.gg`(ggplot(mtcars), aes(wt, mpg))
#>   2.   \-ggplot2:::add_ggplot(e1, e2, e2name) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/plot-construction.r:51:2
#>   3.     +-ggplot2::ggplot_add(object, p, objectname) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/plot-construction.r:69:2
#>   4.     \-ggplot2:::ggplot_add.uneval(object, p, objectname) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/plot-construction.r:87:2
#>   5.       \-ggplot2:::make_labels(object) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/plot-construction.r:126:2
#>   6.         \-base::Map(default_label, names(mapping), mapping) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/aes-calculated.r:104:2
#>   7.           \-base::mapply(FUN = f, ..., SIMPLIFY = FALSE)
#>   8.             \-(function (aesthetic, mapping) ...
#>   9.               \-ggplot2:::strip_dots(mapping) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/aes-calculated.r:95:4
#>  10.                 +-expr[[1]] C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/aes-calculated.r:70:4
#>  11.                 \-rlang:::`[[.quosure`(expr, 1) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/aes-calculated.r:70:4
#>  12.                   +-base::.doTrace(lobstr::cst(), "on entry")
#>  13.                   | \-base::eval.parent(exprObj)
#>  14.                   |   \-base::eval(expr, p)
#>  15.                   |     \-base::eval(expr, p)
#>  16.                   \-lobstr::cst()
#> Tracing `[[.quosure`(expr, 1) on entry 
#>      x
#>   1. \-ggplot2:::`+.gg`(ggplot(mtcars), aes(wt, mpg))
#>   2.   \-ggplot2:::add_ggplot(e1, e2, e2name) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/plot-construction.r:51:2
#>   3.     +-ggplot2::ggplot_add(object, p, objectname) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/plot-construction.r:69:2
#>   4.     \-ggplot2:::ggplot_add.uneval(object, p, objectname) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/plot-construction.r:87:2
#>   5.       \-ggplot2:::make_labels(object) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/plot-construction.r:126:2
#>   6.         \-base::Map(default_label, names(mapping), mapping) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/aes-calculated.r:104:2
#>   7.           \-base::mapply(FUN = f, ..., SIMPLIFY = FALSE)
#>   8.             \-(function (aesthetic, mapping) ...
#>   9.               \-ggplot2:::strip_dots(mapping) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/aes-calculated.r:95:4
#>  10.                 +-expr[[1]] C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/aes-calculated.r:70:4
#>  11.                 \-rlang:::`[[.quosure`(expr, 1) C:/Users/hiroaki-yutani/Documents/repo/R/ggplot2/R/aes-calculated.r:70:4
#>  12.                   +-base::.doTrace(lobstr::cst(), "on entry")
#>  13.                   | \-base::eval.parent(exprObj)
#>  14.                   |   \-base::eval(expr, p)
#>  15.                   |     \-base::eval(expr, p)
#>  16.                   \-lobstr::cst()
expect_identical(class(p$mapping), "uneval")

^{Created on 2019-10-07 by the reprex package (v0.3.0)}

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/