Recently {waldo} dropped {tibble} as a required package, which caused these tests to break for us:
══ Failed tests ════════════════════════════════════════════════════════════════
── Failure (test-deparse.R:571:3): formulas are deparsed (#1169) ───────────────
expr_deparse(~foo) (`actual`) not equal to "<formula>" (`expected`).
`actual`: "<S3: formula>"
`expected`: "<formula>"
── Failure (test-deparse.R:595:3): matrices and arrays are formatted (#383) ────
as_label(mat) (`actual`) not equal to "<int[,1]>" (`expected`).
`actual`: "<int>"
`expected`: "<int[,1]>"
── Failure (test-deparse.R:596:3): matrices and arrays are formatted (#383) ────
expr_deparse(mat) (`actual`) not equal to "<int[,1]: 1L, 2L, 3L>" (`expected`).
`actual`: "<int: 1L, 2L, 3L>"
`expected`: "<int[,1]: 1L, 2L, 3L>"
── Failure (test-deparse.R:599:3): matrices and arrays are formatted (#383) ────
as_label(mat2) (`actual`) not equal to "<int[,2]>" (`expected`).
`actual`: "<int>"
`expected`: "<int[,2]>"
── Failure (test-deparse.R:600:3): matrices and arrays are formatted (#383) ────
expr_deparse(mat2) (`actual`) not equal to "<int[,2]: 1L, 2L, 3L, 4L>" (`expected`).
`actual`: "<int: 1L, 2L, 3L, 4L>"
`expected`: "<int[,2]: 1L, 2L, 3L, 4L>"
── Failure (test-deparse.R:603:3): matrices and arrays are formatted (#383) ────
as_label(arr) (`actual`) not equal to "<int[,1,3]>" (`expected`).
`actual`: "<int>"
`expected`: "<int[,1,3]>"
── Failure (test-deparse.R:604:3): matrices and arrays are formatted (#383) ────
expr_deparse(arr) (`actual`) not equal to "<int[,1,3]: 1L, 2L, 3L>" (`expected`).
`actual`: "<int: 1L, 2L, 3L>"
`expected`: "<int[,1,3]: 1L, 2L, 3L>"
I haven't looked into the underlying code to comment on what might be preferable :)
rlang/tests/testthat/test-deparse.R
Lines 569 to 605 in 38eb82d
Recently {waldo} dropped {tibble} as a required package, which caused these tests to break for us:
There are a few ways we could improve this; the simplest is to use
skip_if_not_installed("tibble")for those tests where {tibble} is required.I haven't looked into the underlying code to comment on what might be preferable :)