I wrote:
> I notice that EquivalenceClass is already marked as no_copy_equal,
> which means that gen_node_support.pl can know that emitting a
> recursive node-copy or node-compare request is a bad idea. What
> do you think of using the patch as it stands, plus a cross-check
> that we don't emit COPY_NODE_FIELD or COMPARE_NODE_FIELD if the
> target node type is no_copy or no_equal?
Concretely, it seems like something like the attached could be
useful, independently of the other change.
regards, tom lane