Re: [HACKERS] Re: pg_dump possible fix, need testers.

* Alfred Perlstein <bright@wintelcom.net> [000124 17:40] wrote:
> * Patrick Welche <prlw1@newn.cam.ac.uk> [000124 16:02] wrote:
> > On Mon, Jan 24, 2000 at 03:49:26PM -0800, Alfred Perlstein wrote:
> > > I just ran the regression tests as best as I know how:
> > > 
> > > ~/pgcvs/pgsql/src/test/regress % gmake runcheck
> > > ~/pgcvs/pgsql/src/test/regress % grep FAIL run_check.out  
> > >            test int2                 ... FAILED
> > >            test int4                 ... FAILED
> > >            test float8               ... FAILED
> > > sequential test geometry             ... FAILED
> > > ~/pgcvs/pgsql/src/test/regress %
> > > 
> > > no int2/int4? yipes!
> > 
> > Not to worry, those will be differences in error message wording, but
> > 
> > > I ran it 10 more times and one time I got:
> > >            test constraints          ... FAILED
> > 
> > What did this error come from? (cf regression.diffs)
> > 
> > > but i got no weird parse errors or anything from the backend.
> > > 
> > > Have you been able to find any weirdness with the fix I posted,
> > > or is this more likely an issue with Patrick Welche's setup?
> > 
> > I'm not sure: on the one hand, that evil join of mine returns the entire
> > contents of a table, and the connection gets confused. Smaller joins work.
> > Maybe it doesn't happen to you because you don't put in such a useless
> > select (What do you want 750440 rows for?) On the other hand vacuum analyze
> > table_name doesn't work for me but obviously does for everyone else, so at
> > least something is wrong with my setup.
> 
> whoa whoa whoa... I just updated my snapshot to today's code and lot
> more seems broken:

because I forgot to clean & rebuild in the regression dir... oops.

Patrick, I'll have a delta shortly that totally backs out my non-blocking
changes for you to test.

-Alfred