Hi I would like to make a light weight language transformation stuff.

By "language transformation" it means transfer language into a unified form, to eliminate synthetic sugar or eliminate useless expressions. One of the example as follow in C:

// before ==================
int a[2][2];
a[0][0] = 1;
// after ===================
int a[4];
a[0] = 1;

What I'm thinking of is to use flex and bison to find target patterns to replace.

However I don't what is the normal way of doing this kind of things, thus I would like to ask and seeking advices.

Thanks a lot!!!!