[HACKERS] bug #1702: nested composite types in plpgsql

[Roman Neuhauser]

Hello, I'm getting this error with the code below (on 8.0.3 like the
other guy in #1702. Is this a hard problem to fix?

Looking at src/pl/plpgsql/src/pl_exec.c for the first time, is it a problem
of make_tuple_from_row() not accounting for nested composite types?

test=# SELECT takes_ct2parts(1, ARRAY['ct2a'], ARRAY['ct1a'], ARRAY['ct1b']);
ERROR:  row not compatible with its own tupdesc
CONTEXT:  PL/pgSQL function "takes_ct2parts" line 18 at perform
test=# SELECT version();
                               version                                
----------------------------------------------------------------------
 PostgreSQL 8.0.3 on i386-portbld-freebsd4.11, compiled by GCC 2.95.4
(1 row)

CREATE TYPE ct1 AS (
 a text,
 b text
);

CREATE TYPE ct2 AS (
 a text,
 b ct1
);

CREATE FUNCTION takes_ct2(ct2)
RETURNS BOOLEAN
RETURNS NULL ON NULL INPUT
VOLATILE
LANGUAGE plpgsql
AS '
  BEGIN
    RETURN TRUE;
  END;
';

CREATE OR REPLACE FUNCTION takes_ct2parts(int4, text[], text[], text[])
RETURNS BOOLEAN
RETURNS NULL ON NULL INPUT
VOLATILE
LANGUAGE plpgsql
AS '
  DECLARE
    _sz   ALIAS FOR $1;
    _ct2a ALIAS FOR $2;
    _ct1a ALIAS FOR $3;
    _ct1b ALIAS FOR $4;
    _ct1  ct1;
    _ct2  ct2;
    _cnt  INTEGER DEFAULT 1;
  BEGIN
    WHILE _cnt <= _sz LOOP
      _ct1.a := _ct1a[_sz];
      _ct1.b := _ct1b[_sz];
      _ct2.a := _ct2a[_sz];
      _ct2.b := _ct1;
      _cnt   := _cnt + 1;
    END LOOP;
    PERFORM takes_ct2(_ct2);
    RETURN TRUE;
  END;
';

-- 
How many Vietnam vets does it take to screw in a light bulb?
You don't know, man.  You don't KNOW.
Cause you weren't THERE.             http://bash.org/?255991

---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend