Extended Static Race Detection for Visual Studio

Multi-threaded programs are known to be prone to data races that are hard to find due to their
non-deterministic occurrence. This thesis implements a novel static analysis mechanism for detecting
low-level data races in the C# programming language. The tool works on the fly within
the Visual Studio integrated development environment using the Roslyn framework. Based on a
novel algorithm, the analysis employs conservative inter-procedural data-flow analysis considering
thread dependency graphs with start/join relations. The result is working prototype with a simple
architecture that efficiently detects potential data race issues by conveniently highlighting them in
the program source code.