www.digitalmars.com         C & C++   DMDScript  

D - contracts bug

reply "Daniel Yokomiso" <daniel_yokomiso yahoo.com.br> writes:
Hi,

    Is this code valid?


int[] fun(int i)
in {
   assert(i > 0);
} out (result) {
   assert(result.length == 10);
} body {
   int[] res = new int[10];
   res[] = i;
   int isZero = (result[0] == 0);
   printf("Is first item zero? %.*s\r\n", cast(char[]) (isZero ? "true" :
"false"));
   result[0] = 0;
   return res;
}
int main() {
    int[] values = fun(2);
    return 0;
}


    I think it should not be, unless result (declared in out section)
function as an implicit variable, like Eiffel defines the Result implicit
variable. This code compiles and run (dmd 0.50). Statement 'int isZero =
(result[0] == 0);' executes fine, but 'result[0] = 0;' raises an 'Access
Violation'.
    Implicit result variable looks like a good idea, but is terrible when
the language provides closures. The following code demonstrates this:

template TCollection(T) {
    class Collection {
        ...
        boolean all(boolean (*predicate)(T)) {
           ...
        }
        ...
    }
    Collection copyFrom(Collection other)
    out (result) {
       assert(result.length == other.length);
       assert(result !== other);

// if result is an implicit variable, the result used in the
// closure represents the closure result, not the copyFrom result
       assert(other.all(fun(item) {return (item in result);});
    } body {

        ...
    }

    Best regards,
    Daniel Yokomiso.

"Those are my principles. If you don't like them I have others."
- Groucho Marx
Nov 28 2002
parent "Walter" <walter digitalmars.com> writes:
"Daniel Yokomiso" <daniel_yokomiso yahoo.com.br> wrote in message
news:as6i38$1f3e$1 digitaldaemon.com...
 Hi,

     Is this code valid?
Yes. result is declared implicitly.
Dec 02 2002