# Strange behaviour of long long and unsigned int

Discussion in 'C Programming' started by luke, Mar 8, 2006.

1. ### lukeGuest

Hi everybody,
please, can someone explain me this behaviour.

I have the following piece of code:

long long ll;
unsigned int i = 2;
ll = -1 * i;
printf("%lld\n", ll);

Why this prints 4294967294 instead of -2?
Luke

luke, Mar 8, 2006

2. ### GowthamGuest

Just type cast -1 to long long.

ll = (long long)-1 * i;

Thats it.

Gowtham, Mar 8, 2006

3. ### GrumbleGuest

luke wrote:
> Hi everybody,
> please, can someone explain me this behaviour.
>
> I have the following piece of code:
>
> long long ll;
> unsigned int i = 2;
> ll = -1 * i;
> printf("%lld\n", ll);
>
> Why this prints 4294967294 instead of -2?

http://c-faq.com/expr/intoverflow1.html

Grumble, Mar 8, 2006
4. ### Keith ThompsonGuest

"luke" <> writes:
> I have the following piece of code:
>
> long long ll;
> unsigned int i = 2;
> ll = -1 * i;
> printf("%lld\n", ll);
>
> Why this prints 4294967294 instead of -2?

In
ll = -1 * i;
the expression -1 is of type int, and i is of type unsigned int, so
the -1 is converted to unsigned int before the multiplication.
Converting -1 to unsigned int yields UINT_MAX, which happens to be
4294967295 in your implementation. Multiplying 4294967295 by 2 wraps
around (because it's unsigned arithmetic), yielding 4294967294.

--
Keith Thompson (The_Other_Keith) <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst>
We must do something. This is something. Therefore, we must do this.

Keith Thompson, Mar 8, 2006
5. ### lukeGuest

Thanks.
Where can I find the rules for arithmetic conversions?
I found this rules (taken from an ANSI C manual):

"First, if the corresponding real type of either operand is long
double, the other
operand is converted, without change of type domain, to a type whose
corresponding real type is long double.
Otherwise, if the corresponding real type of either operand is double,
the other
operand is converted, without change of type domain, to a type whose
corresponding real type is double.
Otherwise, if the corresponding real type of either operand is float,
the other
operand is converted, without change of type domain, to a type whose
corresponding real type is float.
Otherwise, the integer promotions are performed on both operands. Then
the
following rules are applied to the promoted operands:
If both operands have the same type, then no further conversion is
needed.
Otherwise, if both operands have signed integer types or both have
unsigned
integer types, the operand with the type of lesser integer conversion
rank is
converted to the type of the operand with greater rank.
Otherwise, if the operand that has unsigned integer type has rank
greater or
equal to the rank of the type of the other operand, then the operand
with
signed integer type is converted to the type of the operand with
unsigned
integer type.
Otherwise, if the type of the operand with signed integer type can
represent
all of the values of the type of the operand with unsigned integer
type, then
the operand with unsigned integer type is converted to the type of the
operand with signed integer type.
Otherwise, both operands are converted to the unsigned integer type
corresponding to the type of the operand with signed integer type."

Are this rules good or maybe can you suggest anything better?
Thanks

Keith Thompson wrote:

> "luke" <> writes:
> > I have the following piece of code:
> >
> > long long ll;
> > unsigned int i = 2;
> > ll = -1 * i;
> > printf("%lld\n", ll);
> >
> > Why this prints 4294967294 instead of -2?

>
> In
> ll = -1 * i;
> the expression -1 is of type int, and i is of type unsigned int, so
> the -1 is converted to unsigned int before the multiplication.
> Converting -1 to unsigned int yields UINT_MAX, which happens to be
> 4294967295 in your implementation. Multiplying 4294967295 by 2 wraps
> around (because it's unsigned arithmetic), yielding 4294967294.

luke, Mar 8, 2006

luke wrote:
> Thanks.
> Where can I find the rules for arithmetic conversions?

The ISO (and ANSI) C Standard.

> I found this rules (taken from an ANSI C manual):
>
> "First, if the corresponding real type of either operand is long
> double, the other

....

> Are this rules good or maybe can you suggest anything better?

They are correct and useful, assuming you understand what is meant by
integer rank.

--

7. ### lukeGuest

Mmmh, I think rank is what is returned by sizeof() but I'm not sure
Can you explain me what rank really means?
Thanks
Luke

> luke wrote:
> > Thanks.
> > Where can I find the rules for arithmetic conversions?

>
> The ISO (and ANSI) C Standard.
>
> > I found this rules (taken from an ANSI C manual):
> >
> > "First, if the corresponding real type of either operand is long
> > double, the other

> ...
>
> > Are this rules good or maybe can you suggest anything better?

>
> They are correct and useful, assuming you understand what is meant by
> integer rank.
>
> --

luke, Mar 9, 2006
8. ### lukeGuest

In another ANSI C manual I found the following rules, which are
different from the previous ones.
Which one are good then??

1.

If either operand is long double, the other operand is converted
to long double.
2.

If either operand is double, the other operand is converted to
double.
3.

If either operand is float, the other operand is converted to
float.
4.

Integral promotions are performed on both operands, and then the
rules listed below are followed. These rules are a strict extension of
the ANSI "Usual Arithmetic Conversions" rule (Section 3.2.1.5). This
extension ensures that integral expressions will involve long long only
if one of the operands is of type long long. For ANSI conforming
compilation, the integral promotion rule is as defined in Section
3.2.1.1 of the Standard. For non-ANSI compilation, the unsigned
preserving promotion rule is used.
1.

If either operand is unsigned long long, the other operand
is converted to unsigned long long,
2.

otherwise, if one operand is long long, the other operand
is converted to long long,
3.

otherwise, if either operand is unsigned long int, the
other operand is converted to unsigned long int,
4.

otherwise, if one operand is long int, and the other is
unsigned
int, and long int can represent all the values of an
unsigned
int, then the unsigned int is converted to a long int. (If
one operand is long int, and the other is unsigned int, and long int
can NOT represent all the values of an unsigned int, then both operands
are converted to unsigned long int.)
5.

If either operand is long int, the other operand is
converted to long int.
6.

If either operand is unsigned int, the other operand is
converted to unsigned int.
7.

Otherwise, both operands have type int.
Regards

> > Are this rules good or maybe can you suggest anything better?

>
> They are correct and useful, assuming you understand what is meant by
> integer rank.
>
> --

luke, Mar 9, 2006
9. ### Richard BosGuest

"luke" <> wrote:

> Mmmh, I think rank is what is returned by sizeof() but I'm not sure

No, but it's a related concept. sizeof returns the actual storage size
any type or object takes in bytes of memory; rank is a (relative, and
non-measurable) indicator of the ideal mathematical size of an integer
type.

> Can you explain me what rank really means?

It is defined precisely in paragraph 6.3.1.1 of the Standard, but
basically, it's an indicator - not even a number - of how "large" a type
is, in the sense of what it can and must be able to contain. For
example:
- a signed type and its corresponding unsigned type have the same rank
- no signed integer types have the same rank, even if they look the same
in all respects
- a signed type which can hold larger numbers has higher rank than one
which can hold smaller numbers
- even if, say, INT_MAX == LONG_MAX, the rank of long is higher than
that of int, and similar for the other default types
- _Bool has the lowest rank of all
and so on.

The significance of this rank is that when integer types need to be
compared to one another, types with lower rank are generally converted
to types with higher rank.

Richard

Richard Bos, Mar 9, 2006
10. ### Peter Shaggy HaywoodGuest

Groovy hepcat luke was jivin' on 9 Mar 2006 01:49:02 -0800 in
comp.lang.c.
Re: Strange behaviour of long long and unsigned int's a cool scene!
Dig it!

>In another ANSI C manual I found the following rules, which are
>different from the previous ones.

No they're not. They're the same. They're just stated differently.
They amount to the same thing.

>Which one are good then??

Both. But the set of rules you've quoted here (which I've snipped)
lacks a rule for when both operands have the same type. But otherwise
it's fine.

--

Dig the even newer still, yet more improved, sig!