Nice work, but why not clang's AST vs manually inferring the code? #4
Description
I was planning to work on a very similar thing by modifying the AST produced by clang, but I noticed that you analyze the source as it is.
I am possibly going to contribute to this project, but I wanted to understand why not use clang AST instead.
Possibly, your code will break, if someone does some unexpected modifications in the cursed code.
I also wanted to make it even more pythonic by omitting return 0 and similar things.