Static Analysis Methods For Memory Leak Detection: A Survey

Memory leaks are still one of the common bugs in programs written in languages without garbage collectors. There is a list of methods to detect such bugs. One of the approaches is to use static analysis. This paper discusses static analysis methods for memory leak detection. Mainly those methods use two stages. First, they use different types of graphs to find memory leaks faster but with less precision. Then, for the detected cases they use path-sensitive analysis to increase precision. Additionally, we tested available tools on Juliet test suite. We chose false positive rate (FPR) and false negative rate (FNR) as metrics. On average SMOKE and Clang static analysis tools show the best FPR and FNR.