In this folder all the interesting stuff happens.

nc.py is a definition of a node collection. nv can be lazy to, ncl is described in the lazy.py. After parsing clang gives you a collection of nodes. Nc and ncl allow you query them and only when you access the data the actual iteration is performed.

callgraph.py implements the caller graph. This graph is usually a tree. A loop would mean recursive implementations. They are handled in the call graph too. After the callgraph has been built you can pretty print call stacks and the graph too.

pointerutils collect pointers to functions. Sometimes functions are called over pointers to them. So we track pointers as well in case you want to know more than the caller graph can tell you.

clangparser hides the parse folder functionality. You can also find parse file funtion there.

cg utils is a file describing a caller table building process. A caller table is built to make the caller graph building faster. This optimization made sense in our case, because of the large code base for the analysis.

obsdanalysis is the place where OpenBSD kernel analysis is happening. This is an orthogonal concern for the ground elements discussed above. So you can use call graph and node collections even when you have nothing to do with OpenBSD kernel. It was our application only.

(C) Zaur Molotnikov. zmolo@genua.de