On Tue, 3 Jul 2001, Tom Lane wrote:
> Alex Pilosov <alex@pilosoft.com> writes:
> > On Tue, 3 Jul 2001, Tom Lane wrote:
> >> So you have four (soon to be six or seven) different structs that *must*
> >> have the same fields? I don't think that's cleaner than a union ...
>
> > Please see my diffs. Its implemented via #define to declare all common
> > fields.
> > #define RTE_COMMON_FIELDS \
> > NodeTag type; \
> > [etc]
>
> I don't think that technique is cleaner than a union, either ;-).
> The macro definition is a pain in the neck: you have to play games with
> semicolon placement, most tools won't autoindent it nicely, etc etc.
True true. On other hand, unlike union, its automatically typechecked, you
cannot by mistake reference a field you shouldn't be referencing.
Strict typechecking allows one to explicitly declare which type your
function takes if you want, and force errors if you miscast something. I
think discipline is a good thing here...
But really its your call, no point in arguing :)
> But the main point is that I think NodeType = RangeTblEntry with
> a separate subtype field is a better way to go than making a bunch of
> different NodeType values. When most of the fields are common, as in
> this case, it's going to be true that many places only want to know
> "is it a rangetable entry or not?"
That's why I have IsA_RTE(node) macro. (Same as we have IsA_Join and
IsA_JoinPath already). It