python script to match C function signature in multiple lines

Question

I am reading .c file to look out for functions defined in it and count number of lines in each function. My problem is that I am unable to look for function name/signature spanned across multiple ines. I have the list of function names of .c file and i am matching the function names of this list with functions in .c file to process further.

e.g. My .c file is:

 1.   int main(
 2.           void
 3.          )

here main signature is spanned over three lines.

and I have a list of functions as:

 int main(void);

how can i match "int main(void)" with multiline main in .c file ? I want to start counting lines once function is matched.

Answer 1

I suggest you write a simple parser for the C language.

One of the examples in the ANTLR book does something similar to what you're after.

Pyparsing is a very nice Python library for writing parsers.

Here is a parser for ANSI C: http://code.google.com/p/pycparser/ (written using another Python parser library, Ply).

Answer 2

If you have definition that matches "exactly" then you can use regex:

int\s+main\s*\(\s*void\s*\)\s*;

where \s* means zero or more whitechars, and \s+ one or more whitechars.

To use it with multiline search you can define it like:

RE_MAIN = re.compile(r'int\s+main\s*\(\s*void\s*\)\s*;', re.MULTILINE | re.DOTALL)

By "exactly" I mean that it does not match function definition like

int main();

(void omitted)

This way you can find where the function begins, then do simple char scanner counting { and } remembering to ignore comments and ignore chars and strings constants