>Shouldn't that be
>
>long long foo(int a, int b)
>{
>  return (long long)(a*b);
>}
>
>smull taking 32 bit values and giving a 64 out?

Emperically, no -- that calculates a 32-bit product and then sign extends to 
64 bits.  I don't have a copy of the standard on hand to see what it says 
about automatic promotion in cases like this.

p.


unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]

Reply via email to