testClassTimer, testFunctionTimer, testExecTime are not robust
Those tests actually measure speed, which leads to erratic failures. I have tried to increase the margins but this only improved but not solved the problem.
I am wondering, if we either increase the testing-margins even much more to make them robust, or if we find a way to NOT test times, but rather just fake-times, which would be much better. In the end we want to unit-test the interfaces, not the timing accuracy...