| Kurzfassung
|
Dead, unused or unreachable code degrades source code maintainability and comprehensibility. Dead code analyses aim to identify such code, allowing developers to safely remove it. However, due to dynamic language features, purely static analyses often produce false positives. This master's thesis presents a hybrid approach to reduce these false positives. We first analyze the code statically to identify potential dead code. Based on these static results, we then instrument the code to collect runtime information about which parts are actually used during execution. This combination of static analysis with dynamic validation allows us to filter out false positives, providing developers with more accurate and reliable results for improving code health.
|
