Currently a lot of unit tests fail for reasons not entirely clear (or due to the way how the unit test is executed). This is very bad in terms of indication since every unit test failure needs to be inspected manually if it's really a bug in the new patch.