Re: [HACKERS] Error in new psql

Bruce Momjian wrote:

> > actually I meant
> >
> > CREATE RULE myrule AS ON DELETE TO mytable DO (
> > DELETE FROM myothertab1 WHERE key = old.key;
> > DELETE FROM myothertab2 WHERE key = old.key;
> > );
> > ERROR: parser: parse error at or near ""
>
> OK, I fixed it. Just one addition test in an _if_ statement.

Thank you.

You remember, that it's not the first time multiple action
rules have been broken? The other one was due to the
EXCEPT/INTERCEPT patch.

I added a check to the rules regression test after that, to
ensure it never happens again. Unfortunately, Peter's
enforcement to use old psql for regression prevented it from
showing up.

Don't misunderstand this as some whining about it. It is a
very important issue. It shows that the changes made to psql
can cause backward incompatibilities by themself.

AFAIK, the proposed procedure to activate the new psql was to
run the regression test with an old psql, if it's O.K. run it
again with the new one and replace all expected output files.
THIS IS INADEQUATE according to the results seen in this
case.

Don't know if anyone would feel comfortable with it, but at
least, the postmaster log must be checked to show up exactly
the same too. The only alternative would be to check every
old/expected to new/results manually (what's really a whole
lot of damned stupid work).

Jan

--

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