Performance of a SETOF returning function which returns results of another SETOF function?

Question

I have written a function that takes two arguments and returns a SETOF result.

CREATE FUNCTION foo(parentId bigint, childId bigint) RETURNS SETOF bar AS ...

I would like to write two "wrappers" for this function that make it simpler to call:

CREATE FUNCTION foo_parent(parentId bigint) 
RETURNS SETOF bar AS 
...
BEGIN
  RETURN QUERY SELECT * FROM foo(parentId, NULL);
END;

CREATE FUNCTION foo_child(childId bigint) 
RETURNS SETOF bar AS 
...
BEGIN
... look up parent ID ...
RETURN QUERY SELECT * FROM foo(parentId, childId);
END;

Is postgres smart enough to directly return the result of FOO() when FOO_CHILD() is called, or will it copy all the results from FOO() while FOO_CHILD() executes, then return those results?

I think I am asking if pgplsql will do tail-call optimization!

EDIT: I am using Postgres 8.2.

Answer 1

plpgsql always materializes the results before returning, so no.