add 'xfail' to 'LearnerND' for 'BalancingLearner' and link issue

adaptive/tests/test_learners.py

@@ -351,7 +351,8 @@ def test_learner_performance_is_invariant_under_scaling(learner_type, f, learner
     assert abs(learner.loss() - control.loss()) / learner.loss() < 1e-11
 
 
-@run_with(Learner1D, Learner2D, LearnerND, AverageLearner)
+# XXX: The LearnerND shouldn't fail, see https://gitlab.kwant-project.org/qt/adaptive/issues/105
+@run_with(Learner1D, Learner2D, xfail(LearnerND), AverageLearner)
 def test_balancing_learner(learner_type, f, learner_kwargs):
     """Test if the BalancingLearner works with the different types of learners."""
     learners = [learner_type(generate_random_parametrization(f), **learner_kwargs)