Why reference to indirection to pointer is allowed?

Question

int *ptr = 0; int &ref = *ptr;

I write above code in Visual Studio and it works? Here I am pointing to NULL. Why it is allowed? Pointer can take any address, NULL or even invalid address. Still reference to indirection of pointer is allowed?

Then why it is said "Reference cannot be null." Here, is reference not pointing to NULL?

@T.C. Which is undefined, referencing to pointer, or referencing to null pointer? — Pranit Kothari
– Pranit Kothari, Commented Jan 6, 2015 at 16:15
I don't think dereferencing yields a temporary, that's why you can take the address of it. — BWG
– BWG, Commented Jan 6, 2015 at 16:17
"Still pointer to reference is allowed?" – no it isn't. But there's no pointer to reference anywhere in this piece of code. — The Paramagnetic Croissant
– The Paramagnetic Croissant, Commented Jan 6, 2015 at 16:17
@PranitKothari: Dereferencing a null pointer is undefined. That is, the expression *ptr causes undefined behavior. — Bill Lynch
– Bill Lynch, Commented Jan 6, 2015 at 16:18

Mike Seymour · Accepted Answer · 2015-01-06 16:20:46Z

It's not allowed.

In the words of the standard:

C++11 8.3.2/5: A reference shall be initialized to refer to a valid object or function.

However, it's not something that the compiler can, in general, diagnose, since the validity of a pointer is a run-time thing. So there's no requirement to diagnose the error, and probably no compiler warning, just undefined behaviour.

The standard specifically mentions this case:

Note: in particular, a null reference cannot exist in a well-defined program, because the only way to create such a reference would be to bind it to the “object” obtained by dereferencing a null pointer, which causes undefined behavior.

This might be a purely language-laywer remark, but it can be argued that dereferencing a null pointer does not cause undefined behaviour. On the other hand, it seems the committee is opposed to introducing null references. That means, the two are unrelated. See open-std.org/jtc1/sc22/wg21/docs/cwg_active.html#232
@dyp: Indeed, there's enough ambiguity in the standard here to argue that, if you care about that level of pedantry. But it's off-topic here; it's certainly UB to use the result to initialise a reference, since it doesn't refer to a valid object.
Technically the quoted sentence of 8.3.2/5 requires a diagnostic (it's not marked with "no diagnostic required"), which would of course be completely unrealistic.

Collectives™ on Stack Overflow

Why reference to indirection to pointer is allowed?

1 Answer 1

3 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Related