Parsing a C function in Python [closed]

Question

Say I want to parse a C function in a Python script and want to get a list with all the types of the arguments and only the arguments for the C function. How would I go about doing this efficiently?

E.g.: With this function

uint32_t foo(char a, int b, double* c, uint64_t d);

I want output ['char', 'int', 'double*', 'uint64_t']

Answer 1

I think what you want is to parse the "prototype of a C function". When you say parse a "C function" ordinarily one would assume that you are doing what a compiler might do with the source implementation.

In your example, you only show a prototype. In the real world, you will encounter many complications that might make you choose another approach. Your killer is the C preprocessor. In an insane piece of code the real type of 'a' could actually be double* due to a really bad macro. You also might have pointers to structs, typedefs for function pointers, etc.

There is a complication that absolutely defies a solution to your problem. The compilation of the prototype will depend on command line arguments to a C compiler describing where to find header files. Without that metadata, you are hosed. With it, you would really want to run the C preprocessor first - check your compiler documentation for how to do so. Then your python program could read the output.

That line might not even be in it - LOL - if it happened to be surrounded by a #if construct that eliminated it.

These features of the C language are the features that make it really nasty to do things like write programs that operate on C, such as a refactoring engine.