Unit testing is a tool for assessing code quality. Unit tests check the correctness of code fragments like methods, loops and conditional statements. Usually, every code fragment is involved in different tests. We propose a classification of tests, depending on the tested features, which delivers a higher detailed feedback than unclassified tests. Unclassified tests only deliver a feedback whether they failed or succeeded. The detailed feedback from the classified tests help to do a better code quality assessment and can be incorporated in tools helping to improve code quality. We demonstrate the power of this approach doing unit tests on network configuration.