Robert,
* Robert Haas (robertmhaas@gmail.com) wrote:
> On Thu, Jan 25, 2018 at 4:52 PM, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
> > My proposal is that instead of looking at three hundred lines, you'd
> > look for 50 lines of `pg_restore -l` output -- is element XYZ in there
> > or not. Quite a bit simpler for the guy adding a new test. This tests
> > combinations of pg_dump switches: are we dumping the right set of
> > objects.
>
> My counter-proposal is that we remove the test entirely. It looks
> like an unmaintainable and undocumented mess to me, and I doubt
> whether the testing value is sufficient to justify the effort of
> updating it every time anyone wants to change something in pg_dump.
Considering it turned up multiple serious bugs, particularly in the
binary upgrade path, I can't disagree more. If you have a counter
proposal which actually results in better test coverage, that'd be
fantastic, but I wholly reject the notion that we should be considering
reducing our test coverage of pg_dump.
Thanks!
Stephen