bug#65491: [PATCH] Improve performance allocating vectors

[Stefan Monnier]

> Within Standard C, the result of converting a pointer value to an
> integer and vice versa is also implementation defined behavior.

FWIW, last I checked it's literally impossible to implement our
conservative GC (or `malloc` for that matter) without relying on
undefined and implementation defined behaviors in C.

So the best we can do is to try and avoid those undefined behavior that
compilers *do* use to bite in the rear.

I still haven't seen any compiler that tries to make use of the
implementation defined behavior of conversion from pointer to integer as
a basis for optimization, so AFAIK we're still safe using those.

Even converting them back to their original pointer (which is what we do
with tag/untag pairs) is documented to be well-defined if you compile
using GCC.

In contrast the pointer arithmetic on NULL pointers appears to be
something which compilers have started to (ab)use as an assumption for
their optimizations.  Hence the need to update our code.


        Stefan