loop representation in gcc during parsing

loop representation in gcc during parsing

[Isil Dillig]

Hi,

I'm trying to use GCC 4.3.0 for a program analysis frontend, and I've 
noticed that for, while, and do-while loops are all converted to a lower 
level representation using goto statements during the parsing phase. 
I've been trying to recover the source-code-like higher-level loop 
representation; however, I'm confused about a few things.

Consider the following nested for loop:

int i, j;
for(i = 0; i<size; i++)
{
    for(j=i, j>i, j--){
    ....
    }
    ...
}

In the function c_parser_for_statement in the file c-parser.c, it seems 
like the body of the outer for loop is an _empty statement list_. It 
seems to me that this statement list is somehow being correctly patched 
at a later point in time, but I don't quite understand how this works. 
My idea was to maintain a mapping from the original loop representation 
with the original loop body, init-expression, etc to the modified lower 
level gcc representation, but since the body is an empty statement list 
at this point, I can't recover the original structure of the loop.

I'd really appreciate any explanation about how specifically loops are 
converted to a lower-level representation and why the body of the loop 
looks like an empty statement list during parsing. Or if there is a 
better way of recovering the higher-level loop representation, I would 
also appreciate any tips about that.

Thanks a lot in advance,
Isil