I am studying the properties of control flow graphs (CFGs) in general. The tool that I use to statically extract CFGs and label them is based on LLVM. Now, I am looking for a large corpus of C/C++ programs (>= 100 programs) that could be used in experiments. There are basically only two requirements:
- Programs should be easy to compile using
clang. Preferably single file programs or simple
- the set of programs should represent multiple application domains.
Although, at this point, the first requirement is prioritized.