I met a "strange" thing...

I met a "strange" thing...

[Graziano Servizi]

Hi,

by pure chance, during a copy+paste editing operation of a source code,
I ended up inadvertently with a code line like the following:

                          int * x = int();

Well, this code HAD BEEN COMPILED by g++ 4.8.3, and I cannot figure out 
its meaning (from the compiler's point of view).

Here is a code sample of what happened to me:

int main()
  {
   int * x = int();  // compiled
   *x = 888;         // compiled
   std::cout << *x << std::endl;  // compiled, but segfault at execution
   // I can understand the last event ...
  }

What is the semantic of a kind of "default constructor of int" assigned 
to a pointer?

Thanks for your attention.

                        G. Servizi

Re: I met a "strange" thing...

[Jonathan Wakely]

On 10 December 2014 at 13:03, Graziano Servizi wrote:
> Hi,
>
> by pure chance, during a copy+paste editing operation of a source code,
> I ended up inadvertently with a code line like the following:
>
>                          int * x = int();
>
> Well, this code HAD BEEN COMPILED by g++ 4.8.3, and I cannot figure out its
> meaning (from the compiler's point of view).

int() creates an integer with value zero.

In C++03 any integer expression with value zero is a valid null
pointer value, so you can do any of these:

int* p1 = 0;
int* p2 = '\0';
int* p3 = 1-1;

Re: I met a "strange" thing...

[Jonathan Wakely]

On 10 December 2014 at 13:09, Jonathan Wakely wrote:
> On 10 December 2014 at 13:03, Graziano Servizi wrote:
>> Hi,
>>
>> by pure chance, during a copy+paste editing operation of a source code,
>> I ended up inadvertently with a code line like the following:
>>
>>                          int * x = int();
>>
>> Well, this code HAD BEEN COMPILED by g++ 4.8.3, and I cannot figure out its
>> meaning (from the compiler's point of view).
>
> int() creates an integer with value zero.
>
> In C++03 any integer expression with value zero is a valid null
> pointer value, so you can do any of these:
>
> int* p1 = 0;
> int* p2 = '\0';
> int* p3 = 1-1;

I should have said any integer *constant* expression.

The exact wording was:

"A null pointer constant is an integral constant expression (5.19)
prvalue of integer type that evaluates to zero or a prvalue of type
std::nullptr_t."

This was modified by
http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html#903 so
that only integer literals are valid, so now int() and 1-1 are not
valid null pointer constants.