Re: [HACKERS] Index corruption

Tom Lane wrote:

> Adriaan Joubert <a.joubert@albourne.com> writes:
>
> > I saw the message about  lengths in indexes,
> > but howcome this is relevant for procedures?
>
> In 6.5 (and before), there's an index on the prosrc field of pg_proc,
> ie, the definition of the procedure.  There's not any real good reason
> to have such an index, so we've removed it for 7.0 ... but in 6.5 it's
> there and it creates problems if you have long procedure definitions :-(

    The usage of it is only #ifdef'd out!

    It's  a  very old standing FEATURE, that doesn't work anyhow.
    It has to do with tuple set's, and as far as I read the  code
    in  question,  the  (no  longer  supported either) nested dot
    notation looked for a 'sql' language function returning a set
    of tuples and created that on the fly.  Therefore, it checked
    by the required functions source text if it exists.

    IIRC the #ifdef is somewhat like SETS_FIXED.


Jan

--

#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me.                                  #
#========================================= wieck@debis.com (Jan Wieck) #