Re: patch fixing the old RETURN NEXT bug

Added to TODO:

       o Fix problems with RETURN NEXT on tables with
         dropped/added columns after function creation

          http://archives.postgresql.org/pgsql-patches/2006-02/msg00165$


---------------------------------------------------------------------------

Sergey E. Koposov wrote:
> Hello All,
>
> I'm proposing the fix of this bug:
> http://archives.postgresql.org/pgsql-hackers/2005-02/msg00498.php
>
> The exact SQL code exposing the error:
>
> ----------
>
> create table usno (ra real, dec real, bmag real, rmag real,ipix int8);
> CREATE OR REPLACE FUNCTION xxx(refcursor) RETURNS refcursor AS '
> DECLARE query varchar;
> BEGIN
> query = ''SELECT * FROM usno'';
> OPEN $1 FOR EXECUTE query;
> RETURN $1;
> END;
> ' LANGUAGE plpgsql;
>
> CREATE OR REPLACE FUNCTION yyy() RETURNS SETOF usno AS '
> DECLARE rec record;
> DECLARE cur refcursor;
> BEGIN
> cur=xxx(''curs_name'');
> LOOP
>         FETCH cur into rec;
>         EXIT WHEN NOT FOUND;
>         RETURN NEXT rec;
> END LOOP;
> RETURN;
> END;
> ' LANGUAGE plpgsql;
>
> insert into usno values(1,2,3,4);
> select * from yyy();
> alter table usno add column errbox box;
> select * from yyy();
> alter table usno drop column errbox;
> select * from yyy();
>
> -------
>
> The problem with that is in fact in pl_exec.c in function
> compatible_tupdesc(), which do not check for the deleted attributes.
>
> The patch is attached.
>
> Regards,
>     Sergey
>
> *****************************************************
> Sergey E. Koposov
> Max Planck Institute for Astronomy
> Web: http://lnfm1.sai.msu.ru/~math
> E-mail: math@sai.msu.ru
>

Content-Description:

[ Attachment, skipping... ]

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

--
  Bruce Momjian   http://candle.pha.pa.us
  EnterpriseDB    http://www.enterprisedb.com

  + If your life is a hard drive, Christ can be your backup. +