On: GCC null pointer check removed

[a1369209993]

> aggressive ["]optimization[" ...] exploiting undefined behavior

That's still a compiler bug.

But yes, this is a bug in GCC and the title should show that.

[saagarjha → a1369209993]

Exploiting undefined behavior is something compilers can and will do; that's not a bug. (Specifically, if a compiler can choose to omit code where it can prove that going down that branch would lead to undefined behavior. That's not what happened here.)

[paulddraper → a1369209993]

To be clear, it's a bug for code to exploit undefined behavior.

A compiler may exploit the assumption that source code does not invoke undefined behavior.

[Skunkleton → a1369209993]

Comparing a pointer to null isn't UB. De-referencing invalid pointers is ub.