Stephen Frost <> writes:
> Indeed, but code like this makes me wonder if this is really working the
> way it's supposed to:

> +   val1 = (long)pk_left->pk_eclass;
> +   val2 = (long)pk_right->pk_eclass;

Ugh.  Casting a pointer to long?  We do have portable ways to do what
this is trying to do, but that is not one.  (For example, this is
guaranteed to misbehave on 64-bit Windows.)

Offhand I think PointerGetDatum might be the best way.

                        regards, tom lane

Sent via pgsql-hackers mailing list (
To make changes to your subscription:

Reply via email to