Test failures on cran

[mb706]

https://www.r-project.org/nosvn/R.check/r-devel-linux-x86_64-fedora-gcc/mlr3pipelines-00check.html

 ── Failure ('test_pipeop_learnercv.R:124:3'): PipeOpLearnerCV - cv ensemble averages fold learners ──
 Expected `all.equal(object, expected, check.environment = FALSE, ...)` to be TRUE.
 Differences:
 `actual` is a character vector ('4 string mismatches')
 `expected` is a logical vector (TRUE)
 
 Backtrace:
     ▆
  1. └─global expect_equal(as.character(result_response), task$class_names[max.col(manual_prob)]) at test_pipeop_learnercv.R:124:3
  2.   └─testthat::expect_true(...)

mlr3pipelines/tests/testthat/test_pipeop_learnercv.R

Lines 124 to 127 in cb8265b

	expect_equal(
	as.character(result_response),
	task$class_names[max.col(manual_prob)]
	)

[advieser]

The reason for the string mismatches is that in these cases the probabilities for classes are tied and both methods (LearnerCV and the manual prediction) resolve those independent of each other, manual predictions with ties.method = "random", and LearnerCV with "first".