Re: [Ocaml-lib-devel] Re: [Caml-list] looping recursion

[skaller <sk...@us...>]

On Fri, 2004-07-30 at 02:04, Brian Hurt wrote:
> On 29 Jul 2004, skaller wrote:

> > Before you laugh too hard at this example -- 
> > THE most common lexer and parser combination
> > in the world must do precisely what I described,
> > its called a C compiler :)
> 
> Actually, in the case of C this isn't true.  There is enough information 
> in the grammar to be able to differentiate type names from variable names.  

I'm afraid not: counter-example: (A)(B)(C)

Here's an actual example, which compiles and runs
fine under gcc:

int f(int x){ return 1; }

typedef int (*fp)(int);

fp g(fp x){ return x; }

int main()
{
  int x = (int)(f)(1); // parsed as (int)(f(1))
  int y = (g)(f)(1);   // parsed as (g(f))(1)
  printf("%d, %d\n", x,y);
}

Application has a higher precedence than casts in C:)

There are other cases which are merely LR(1) ambiguous
but which can be parsed correctly, as you claim,
provided you can look ahead to the end of statement.


-- 
John Skaller, mailto:sk...@us...
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net