Re: [HACKERS] pg_dump not in very good shape

Tom Lane wrote:

> Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> > Actually the megatest is:
>
> > pg_dump regress > out
> > dropdb regression
> > createdb regression
> > psql regression < out
> > pg_dump regress > out2
> > diff out out2
>
> > That is the pg_dump test, and someone usually does it as part of
> > regression testing before each release.
>
> > It would be nice to add this to test/regress/Makefile maybe.
>
> That's a good thought --- it eliminates both the platform-specific
> issues and the problem of adding a bulky reference file to the
> distribution.

Still an incomplete test at all.

It doesn't guarantee, that the resulting dump is what you
really need to restore the database. For example, I'm not
sure that FOREIGN KEY constraints are actually dumped
correct (as constraint trigger statements after the data
load). So it might work, and result in the same, wrong dump
again.

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) #